SBML Discussion Lists RDF feed

Thinking about changing COBRA Toolbox approach to SBML

Ben Heavner — 2013-06-11T18:41:25-00:00

Hi all - I'm revisiting the COBRA-specific use of notes fields, which I think most people here are familiar with. In short, the COBRA Toolbox is software that gained traction among people doing constraint-based analysis of metabolic network models with MATLAB. As documented in supplemental file S.1 of 10.1038/nprot.2011.308, version 2.05 of the COBRA Toolbox uses SBML level 2 version 4 to exchange models. However, it uses a particular key:value mapping to put reaction annotation information in notes fields, like this (from the iND750 model of the yeast metabolic network): Abbreviation: R_13BGH</html:p> Synonyms: _1</html:p> EC Number: 3.2.1.58</html:p> SUBSYSTEM: Starch and Sucrose Metabolism</html:p> Equation: [c] : 13BDglcn + h2o --> glc-D</html:p> Confidence Level: 0</html:p> NCD</html:p> genes:</html:p> LOCUS:YGR282C#ABBREVIATION:BGL2#ECNUMBERS:3.2.1.58#</html:p> proteins:</html:p> NAME:Endo-beta-1,3-glucanase of the cell wall#ABBREVIATION:Bgl2-e#</html:p> citations:</html:p> none</html:p> GENE ASSOCIATION: (YGR282C)</html:p> </notes> At the moment, the COBRA Toolbox as distributed uses the following keys: GENE_ASSOCIATION GENE ASSOCIATION SUBSYSTEM EC Number FORMULA CHARGE AUTHORS Confidence Over time, a variety of groups have published models that are usable with the COBRA toolbox to various degrees, and the list of keys these models use has expanded (though the COBRA Toolbox doesn't support these other keys at the moment - there may be other metabolic network reconstruction software that does). Almost every model includes other keys that the model writers have thought important (as you can see from the iND750 example above). Other keys in notes fields include things like: Abbreviation Synonyms Equation NCD genes Locus ECNUMBERS proteins NAME ABBREVIATION citations INCHI EHMN_ABBREVIATION HEPATONET_1.0_ABBREVIATION At the moment, I'm working to refactor some of the COBRA Toolbox SBML code, and would like to eventually add support for SBML level 3 + the fbc package. If I am reading it correctly, the level 3, Version 1 SBML spec suggests that the legacy approach isn't the best use of the notes field, which is "intended to serve as a place for storing optional information intended to be seen by humans". Perhaps most of the information currently in these notes fields would be more appropriate in the annotation fields - "optional software-generated content not meant to be shown to humans". As I try to work out a roadmap to gradually transition the software to support legacy models by reading them, and write improved SBML that can support more and more SBML features as model developers adopt them, I want to make sure I'm thinking about the annotation and notes fields correctly. I'm considering an approach of moving these key:value pairs from the notes field to the annotation field. I have two key questions: 1) does this approach break the admonition that "it is critical that data essential to a model definition or that can be encoded in existing SBML elements is not stored in annotations." ? 2) could (should?) such an approach work alongside the MIRIAM rdf:Bag approach described at <a href="http://sbml.org/Community/Wiki/About_annotations_in_Level_2" target="_blank">http://sbml.org/Community/Wiki/About_annotations_in_Level_2</a> by adding another top-level element to the annotation field using a model-specific namespace? For example, does this seem like a good way to go (content is made-up, so these are not the correct genes or other annotation): </rdf:Bag> </bqbiol:isVersionOf> </rdf:Bag> </bqbiol:is> </rdf:Description> </rdf:RDF> </annotation> ... </reaction> In this example, I recongize that information in thismodel:reaction may be redundant with information in the rdf annotation, and would hope that MIRIAM support and parsing in the COBRA Toolbox can improve through efforts to move things from these legacy fields to the appropriate ref:resource links. Note that I haven't discussed this with the COBRA developers, and have no idea if other constraint-based model development software would come around to this approach. At the moment, I'm just trying to come up with an approach that strikes me as a good way to go - and moving the custom "notes" info to custom "annotation" info to free up notes for user-readable free text seems like it may be a good step. What do you all think? -Ben

Infix to MathML parsing

Sarah Keating — 2013-06-06T15:52:52-00:00

Hi People who use libSBML to parse between MathML and Infix Since this has come up again I would like to clarify the current position and future plans. readMathMLFromString writeMathMLToString functions are helper functions for dealing with MathML and are not used as infix-mathml-infix conversions. The strings mentioned here are xml strings representing the MathML. The functions for infix/mathml parsing are char * SBML_formulaToString(ASTNode_t) and ASTNode_t * SBML_parseFormula(const char *). These were originally written specifically to aid the support of SBML L2 which introduced MathML and allow easy conversion of models between L1 (that used only infix) and L2. In fact when they were originally written it never occurred to us that people would use this functionality within their own software to display infix or allow users to enter infix. :-) Now this functionality imposed certain constraints based on the infix functions that were defined in SBML L1; for example 'log' was interpreted as natural log. We have acknowledged that this was becoming a problem to people but given the 'historical' necessity of the existing functions we could not just change their functionality. Last year Lucian produced a new function ASTNode_t* SBML_parseL3Formula(const char *) and a sister function ASTNode_t* SBML_parseL3FormulaWithSettings(const char*, L3ParserSettings_t*) that allowed much more freedom with the infix formula that could be parsed. The optional settings options also allowed the user to specify particular functionality such as selecting to parse 'log' as natural log or log base 10. However we have not yet added the reciprocal function SBML_formulaToL3String that will allow the roundtripping of the strings produced by this new function. If roundtripping is an issue for you, at present it is probably safer to use the original parsing functions rather than mix the L3 variant with the older formulaToString function. It has become necessary to redesign the ASTNode class to allow extensions of the math elements allowed by SBML L3 packages. This redesign will automatically facilitate the new SBML_formulaToL3String function and at one point we decided that it was unnecessary to duplicate effort by adding this functionality to work with the original ASTNode code base; since once the AST redesign is complete this code would become obsolete. However, since this is a highly requested function we will re-evaluate our time allocation and consider how we can introduce the function sooner rather than later. Sarah ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Log and Log10 in SBML/libSBML

Sarah Keating — 2013-06-06T15:47:02-00:00

Hi Herbert The readMathMLFromString writeMathMLToString functions are really helper functions for dealing with MathML and are especially helpful for debugging code BUT are not intended to be used as infix-mathml-infix conversions. In fact the strings referred to in these functions will be of xml strings of the actual MathML format. You would use these functions if you are dealing with xml strings such as " ..." Sarah ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Log and Log10 in SBML/libSBML

Lucian Smith — 2013-06-05T05:44:59-00:00

* hsauro <<a href="mailto:hsauro@kgi.edu" target="_blank">hsauro@kgi.edu</a>> [2013-06-05 05:55] writes: > > Just realized, is there are C API for changing the settings? I can't call C++ method through the dll interface, > > l3ps.setParseLog(L3P_PARSE_LOG_AS_LN); Yup; use L3ParserSettings_setParseLog() <a href="http://sbml.org/Software/libSBML/docs/cpp-api/_l3_parser_settings_8cpp.html#a83c9f03cd831628981380b5173d7f919" target="_blank">http://sbml.org/Software/libSBML/docs/cpp-api/_l3_parser_settings_8cpp.html#a83c9f03cd831628981380b5173d7f919</a> -Lucian ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Log and Log10 in SBML/libSBML

hsauro — 2013-06-05T05:10:01-00:00

Lucian, I figured it out, I found after some searching the web the C API used in some code written by Sarah. It seems to work. I have some old code that uses the calls: readMathMLFromString() writeMathMLToString() What are these? Should I be using them? I think I am now into my third version of the libsbml infix/mathml converters.

Re: Log and Log10 in SBML/libSBML

hsauro — 2013-06-05T04:44:53-00:00

Just realized, is there are C API for changing the settings? I can't call C++ method through the dll interface, l3ps.setParseLog(L3P_PARSE_LOG_AS_LN);

Re: Log and Log10 in SBML/libSBML

hsauro — 2013-06-05T02:10:22-00:00

Thank you! I';; give it a go.

Re: Log and Log10 in SBML/libSBML

Lucian Smith — 2013-06-04T23:30:12-00:00

* hsauro <<a href="mailto:hsauro@kgi.edu" target="_blank">hsauro@kgi.edu</a>> [2013-06-05 00:24] writes: > > Thanks for the explanation. How do I use the settings variant because the L3ParserSettings_t seem undocumented? Do you have a code fragment you can provide? Hmm, that should have sent you to <a href="http://sbml.org/Software/libSBML/docs/cpp-api/class_l3_parser_settings.html" target="_blank">http://sbml.org/Software/libSBML/docs/cpp-api/class_l3_parser_settings.html</a> but no matter. Here's what I do in Antimony: L3ParserSettings l3ps; l3ps.setParseCollapseMinus(true); l3ps.setParseLog(L3P_PARSE_LOG_AS_LN); ASTNode* rootnode = SBML_parseL3FormulaWithSettings(formula.c_str(), &l3ps); (The 'collapse minus' thing is to convert things like '---x' to just '-x'.) -Lucian ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Log and Log10 in SBML/libSBML

hsauro — 2013-06-04T23:23:07-00:00

Thanks for the explanation. How do I use the settings variant because the L3ParserSettings_t seem undocumented? Do you have a code fragment you can provide?

Re: Log and Log10 in SBML/libSBML

Lucian Smith — 2013-06-04T22:40:10-00:00

* hsauro <<a href="mailto:hsauro@kgi.edu" target="_blank">hsauro@kgi.edu</a>> [2013-06-04 22:19] writes: > > Something I don't quite understand about logs in SBML. The docs say that (<a href="http://sbml.org/Software/libSBML/docs/java-api/org/sbml/libsbml/libsbml.html" target="_blank">http://sbml.org/Software/libSBML/docs/java-api/org/sbml/libsbml/libsbml.html</a>), the symbol log represents the natural logarithm and log10, log to base 10. > > If I feed the following equation: > > k*log(Node0) > > into the infix to mathml converter (L3, libsbml 5.8) I get the log10 version: > > <?xml version="1.0" encoding="UTF-8"?> > > > > k </ci> > > > > 10 </cn> > </logbase> > Node0 </ci> > </apply></apply> > </math> > > So when I convert it back the natural log has been converted to log10. How do I get the original natural log back? Is there some setting in libsbml or I am using the converters incorrectly? > > The code I am using is (where mathStr is infix): > > astTree = SBML_parseL3Formula (mathStr); > > mathMLStr = writeMathMLToString (astTree); The problem is that while the 'paresL3Formula' can be set to read in the text "log(x)" as either a natural log or a base-10 log (or as an error!), there is not yet a corresponding 'writeMathMLToL3String' function that can also be set the same way. Hence, 'writeMathMLToString' will always write out "log" when it means the natural log, and "log10" when it means the base-10 log. If you want to be able to round-trip infix to XML and back, you'll need to set the parseL3Formula to interpret "log" as "natural log" The easiest way to do this is to use the 'SBML_parseL3FormulaWithSettings' function; see <a href="http://sbml.org/Software/libSBML/docs/cpp-api/_l3_parser_8h.html#a77c72a78afff728a6c569790338f2085" target="_blank">http://sbml.org/Software/libSBML/docs/cpp-api/_l3_parser_8h.html#a77c72a78afff728a6c569790338f2085</a> and <a href="http://sbml.org/Software/libSBML/docs/cpp-api/class_l3_parser_settings.html" target="_blank">http://sbml.org/Software/libSBML/docs/cpp-api/class_l3_parser_settings.html</a> for more details. As soon as Sarah's new ASTNode class setup is in, I can write the 'writeMathMLToL3String' function, and this will become a lot less mysterious (and functional!) -Lucian ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Log and Log10 in SBML/libSBML

hsauro — 2013-06-04T21:02:49-00:00

Something I don't quite understand about logs in SBML. The docs say that (<a href="http://sbml.org/Software/libSBML/docs/java-api/org/sbml/libsbml/libsbml.html" target="_blank">http://sbml.org/Software/libSBML/docs/java-api/org/sbml/libsbml/libsbml.html</a>), the symbol log represents the natural logarithm and log10, log to base 10. If I feed the following equation: k*log(Node0) into the infix to mathml converter (L3, libsbml 5.<img src="images/smiley_icons/icon_cool.gif" border=0 alt="Cool"> I get the log10 version: <?xml version="1.0" encoding="UTF-8"?> k </ci> 10 </cn> </logbase> Node0 </ci> </apply></apply> </math> So when I convert it back the natural log has been converted to log10. How do I get the original natural log back? Is there some setting in libsbml or I am using the converters incorrectly? The code I am using is (where mathStr is infix): astTree = SBML_parseL3Formula (mathStr); mathMLStr = writeMathMLToString (astTree);

Re: [sbml-team] libSBML installation problem and forum access

Frank Bergmann — 2013-05-24T02:32:50-00:00

Hello Christine, I'm afraid I'm not sure what the issue might be. It is not with libSBML, but rather the cmake installation on your machine. You could configure it from the command line. If you let us know what options you need we could give them to you. all the best Frank On May 23, 2013, at 6:58 PM, Christine Leon wrote: > To whom it may concern: > > I downloaded libSBML-5.8.0 and am having trouble with the gui when I click > configure. It crashes and the following error message is displayed in my > terminal. > > Qt has caught an exception thrown from an event handler. Throwing > exceptions from an event handler is not supported in Qt. You must > reimplement QApplication::notify() and catch all exceptions there. > > Do you know what I might be doing incorrectly or can I have access to post > on the forum for help with this issue. > > Thanks so much for your help! > > Christine ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

libSBML installation problem and forum access

Christine Leon — 2013-05-23T22:58:36-00:00

To whom it may concern: I downloaded libSBML-5.8.0 and am having trouble with the gui when I click configure. It crashes and the following error message is displayed in my terminal. Qt has caught an exception thrown from an event handler. Throwing exceptions from an event handler is not supported in Qt. You must reimplement QApplication::notify() and catch all exceptions there. Do you know what I might be doing incorrectly or can I have access to post on the forum for help with this issue. Thanks so much for your help! Christine ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Issues with libsbml matlab binding installation in Ubuntu

chiara — 2013-05-15T10:00:48-00:00

Hello everyone, I'm trying to install the libsbml bindings for MATLAB on my Ubuntu 12.04 machine, but I'm encountering some problems. My Matlab version is 7.14 (Linux 32 bit), and the libsbml version I'm trying to install is 5.8.0. Everything seems to proceed smoothly until I make install the libsbml package, where I get the following error: <pre>make[2]: Leaving directory `/home/chiara/Programmi_SystemsBiology/libsbml-5.8.0/src/sbml' make -C bindings -w install make[2]: Entering directory `/home/chiara/Programmi_SystemsBiology/libsbml-5.8.0/src/bindings' make -C matlab -w install make[3]: Entering directory `/home/chiara/Programmi_SystemsBiology/libsbml-5.8.0/src/bindings/matlab' Failed to determine the mex extension. This can happen if MATLAB cannot be started. (This can be caused by license issues or a missing VPN connection.) Please start MATLAB manually and run buildSBML.m (in src/bindings/matlab). make[3]: *** [all] Error 1 make[3]: Leaving directory `/home/chiara/Programmi_SystemsBiology/libsbml-5.8.0/src/bindings/matlab' make[2]: *** [matlab-recurse] Error 2 make[2]: Leaving directory `/home/chiara/Programmi_SystemsBiology/libsbml-5.8.0/src/bindings' make[1]: *** [bindings-recurse] Error 2 make[1]: Leaving directory `/home/chiara/Programmi_SystemsBiology/libsbml-5.8.0/src' make: *** [src-recurse] Error 2 make: Leaving directory `/home/chiara/Programmi_SystemsBiology/libsbml-5.8.0' </pre> So I move to Matlab and run buildSBML.m, and the output that I get is the following: <pre>>> buildSBML() * Doing preliminary checks of runtime environment ... - This appears to be MATLAB and not Octave. - MATLAB reports the OS is 32-bit Linux. Constructing the libSBML MATLAB interface. * Trying to establish our location ... - We are in the libSBML subdirectory for Matlab. - We appear to be in the libSBML source tree. - We have write access here! That makes us happy. * Locating libSBML library and include files ... - Found /home/chiara/Programmi_SystemsBiology/libsbml-5.8.0/src/.libs/libsbml.so - Root of includes is /home/chiara/Programmi_SystemsBiology/libsbml-5.8.0/src * Creating mex files in /home/chiara/Programmi_SystemsBiology/libsbml-5.8.0/src/bindings/matlab - Building TranslateSBML ... Warning: You are using gcc version "4.6.3-1ubuntu5)". The version currently supported with MEX is "4.4.6". For a list of currently supported compilers see: http://www.mathworks.com/support/compilers/current_release/ - Building OutputSBML ... Warning: You are using gcc version "4.6.3-1ubuntu5)". The version currently supported with MEX is "4.4.6". For a list of currently supported compilers see: http://www.mathworks.com/support/compilers/current_release/ OutputSBML.c: In function â€˜mexFunctionâ€™: OutputSBML.c:193:14: warning: assignment discards â€˜constâ€™ qualifier from pointer target type [enabled by default] Successfully finished. If appropriate, please run "installSBML" next. >> installSBML() Installing the libSBML interface. * Doing preliminary checks of runtime environment ... - This appears to be MATLAB and not Octave. Checking for executables in /usr/local/lib Executables not found </pre> I thought that maybe the problem is that I've installed Matlab in a non-conventional directory, but nevertheless I cannot figure how to move on from this point... Thank you for your help! Chiara

Re: Installing libsbml on MacOSX

fbergman — 2013-04-24T18:13:40-00:00

Hello Johann, the binary is now available below, it needs to be extracted into the site packages directory (/Library/Frameworks/Python.framework/Versions/7.3/lib/python2.7/site-packages) if you don't want to set up the path manually every time: <a href="https://www.dropbox.com/s/x7mm6k98arullcp/2013-04-24_-_libsbml_osx_enthought_7.3-32bit.zip" target="_blank">https://www.dropbox.com/s/x7mm6k98arullcp/2013-04-24_-_libsbml_osx_enthought_7.3-32bit.zip</a> just for the archive, the configure line I used was: cmake -DWITH_PYTHON=ON -DCMAKE_OSX_ARCHITECTURES=i386 -DPYTHON_EXECUTABLE=/Library/Frameworks/Python.framework/Versions/7.3/bin/python -DPYTHON_INCLUDE_DIR=/Library/Frameworks/Python.framework/Versions/7.3/include/python2.7/ -DPYTHON_LIBRARY=/Library/Frameworks/Python.framework/Versions/7.3/Python -DCMAKE_INSTALL_PREFIX=/Users/fbergmann/Development/enthought_install ../libsbml make make install then I was able to use it in enthought: <a href="mailto:fbergmann@setsuna" target="_blank">fbergmann@setsuna</a> ~ $ /Library/Frameworks/Python.framework/Versions/7.3/bin/python Enthought Python Distribution (EPD) free version -- www.enthought.com Version: 7.3-2 (32-bit) (type 'upgrade' or see www.enthought.com/epd/upgrade to get the full EPD) Python 2.7.3 |EPD_free 7.3-2 (32-bit)| (default, Apr 12 2012, 11:28:34) [GCC 4.0.1 (Apple Inc. build 5493)] on darwin Type "credits", "demo" or "enthought" for more information. >>> import sys >>> sys.path.append('/Users/fbergmann/Development/enthought_install/lib/python2.7/site-packages/libsbml/') >>> import libsbml >>> print libsbml.getLibSBMLDottedVersion() 5.8.0 >>> best Frank On Apr 24, 2013, at 4:26 PM, Johann Rohwer <<a href="mailto:jr@sun.ac.za" target="_blank">jr@sun.ac.za</a>> wrote: > Hi Frank > > That would be really great if you could compile them for me. I have *zero* > experience in compiling things on a Mac. > Enthought Python Distribution (EPD) free version -- www.enthought.com > Version: 7.3-2 (32-bit) > > Thanks! > Johann > > On Wednesday 24 April 2013 16:02:59 Frank T. Bergmann wrote: >> Hello Johann, >> >> Enthought does indeed use a different python than the one of the operating >> system. Since python bindings are not compatible among the different >> versions the crash is expected. >> >> So yes, you would have to compile them yourself (or let us do it). I would >> choose CMake, with the configure options: >> >> WITH_PYTHON=ON >> PYTHON_EXECUTABLE=enthought python executable >> PYTHON_INCLUDE_DIR=enthought python include path >> PYTHON_LIBRARY=enthought python library >> CMAKE_OSX_ARCHITECTURES= same as `file enthought python library` >> >> >> I'm not in front of a mac today, but could produce some binaries for you >> tomorrow, which version of EPD are you using? >> >> Thanks >> Best >> Frank >> >>> -----Original Message----- >>> From: <a href="mailto:sbml-interoperability-bounces@caltech.edu" target="_blank">sbml-interoperability-bounces@caltech.edu</a> [mailto:sbml- >>> <a href="mailto:interoperability-bounces@caltech.edu]" target="_blank">interoperability-bounces@caltech.edu]</a> On Behalf Of Johann Rohwer >>> Sent: Wednesday, April 24, 2013 3:37 PM >>> To: <a href="mailto:sbml-interoperability@caltech.edu" target="_blank">sbml-interoperability@caltech.edu</a> >>> Subject: [sbml-interoperability] Installing libsbml on MacOSX >>> >>> I've been trying to install libsbml on MacOSX using the provided stable >> >> 5.8.0 >> >>> binary package installers. I need the python bindings of libsbml. >> >> Importing >> >>> libsbml from a system python prompt works fine. However, because we are >>> interfacing with other python scientific computing software, we are using >> >> the >> >>> Entought Python Distribution (EPD) on these boxes. This is therefore a >>> different python tree from the system python one. I can modify the python >>> path so that the libsbml library files are found from within the EPD >> >> python >> >>> tree, but when I try to import libsbml from the EPD python prompt, it >> >> crashes >> >>> with a huge stack dump. >>> >>> Are the precompiled binaries of libsbml only for the system python on OSX? >>> Does this mean that I have to recompile from sources and link to the EPD >>> python at compile time using the > --with-python and --with-python- >>> interpreter flags? >>> >>> I should add that I am a relative newbie to MacOSX related things but >>> completely at home compiling things in Linux (the Macs are used in a >>> teaching lab that we are setting up). So the concept of different python >>> frameworks on the same machine is a bit foreign to me. >>> >>> Any pointers or ideas would be appreciated... >>> >>> Johann >>> >>> >>> -- >>> Johann M. Rohwer >>> Professor of Systems Biology >>> Dept. of Biochemistry, Stellenbosch University Private Bag X1, 7602 >>> Matieland SOUTH AFRICA >>> Tel +27-21-8085843 Fax +27-21-8085863 >>> <a href="mailto:jr@sun.ac.za" target="_blank">jr@sun.ac.za</a> <a href="http://www.jjj.sun.ac.za" target="_blank">http://www.jjj.sun.ac.za</a> >>> -- >>> Registered Linux User #434886 >>> -- >>> E-pos vrywaringsklousule >>> >>> Hierdie e-pos mag vertroulike inligting bevat en mag regtens >> >> geprivilegeerd >> >>> wees en is slegs bedoel vir die persoon aan wie dit geadresseer is. Indien >> >> u >> >>> nie die bedoelde ontvanger is nie, word u hiermee in kennis gestel dat u >>> hierdie dokument geensins mag gebruik, versprei of kopieer nie. Stel ook >>> asseblief die sender onmiddellik per telefoon in kennis en vee die e-pos >> >> uit. >> >>> Die Universiteit aanvaar nie aanspreeklikheid vir enige skade, verlies of >>> uitgawe wat voortspruit uit hierdie e-pos en/of die oopmaak van enige lC*s >>> aangeheg by hierdie e-pos nie. >>> >>> E-mail disclaimer >>> >>> This e-mail may contain confidential information and may be legally >> >> privileged >> >>> and is intended only for the person to whom it is addressed. If you are >> >> not >> >>> the intended recipient, you are notified that you may not use, distribute >> >> or >> >>> copy this document in any manner whatsoever. Kindly also notify the sender >>> immediately by telephone, and delete the e-mail. The University does not >>> accept liability for any damage, loss or expense arising from this e-mail >>> and/or accessing any files attached to this e-mail. >> >> ____________________________________________________________ >> To manage your sbml-interoperability list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >> >> For a web interface to the sbml-interoperability mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-interoperability list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > -- > Johann M. Rohwer > Professor of Systems Biology > Dept. of Biochemistry, Stellenbosch University > Private Bag X1, 7602 Matieland > SOUTH AFRICA > Tel +27-21-8085843 Fax +27-21-8085863 > <a href="mailto:jr@sun.ac.za" target="_blank">jr@sun.ac.za</a> <a href="http://www.jjj.sun.ac.za" target="_blank">http://www.jjj.sun.ac.za</a> > -- > Registered Linux User #434886 > -- > E-pos vrywaringsklousule > > Hierdie e-pos mag vertroulike inligting bevat en mag regtens geprivilegeerd wees en is slegs bedoel vir die persoon aan wie dit geadresseer is. Indien u nie die bedoelde ontvanger is nie, word u hiermee in kennis gestel dat u hierdie dokument geensins mag gebruik, versprei of kopieer nie. Stel ook asseblief die sender onmiddellik per telefoon in kennis en vee die e-pos uit. Die Universiteit aanvaar nie aanspreeklikheid vir enige skade, verlies of uitgawe wat voortspruit uit hierdie e-pos en/of die oopmaak van enige lês aangeheg by hierdie e-pos nie. > > E-mail disclaimer > > This e-mail may contain confidential information and may be legally privileged and is intended only for the person to whom it is addressed. If you are not the intended recipient, you are notified that you may not use, distribute or copy this document in any manner whatsoever. Kindly also notify the sender immediately by telephone, and delete the e-mail. The University does not accept liability for any damage, loss or expense arising from this e-mail and/or accessing any files attached to this e-mail. ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Installing libsbml on MacOSX

jrohwer — 2013-04-24T14:26:25-00:00

Hi Frank That would be really great if you could compile them for me. I have *zero* experience in compiling things on a Mac. Enthought Python Distribution (EPD) free version -- www.enthought.com Version: 7.3-2 (32-bit) Thanks! Johann On Wednesday 24 April 2013 16:02:59 Frank T. Bergmann wrote: > Hello Johann, > > Enthought does indeed use a different python than the one of the operating > system. Since python bindings are not compatible among the different > versions the crash is expected. > > So yes, you would have to compile them yourself (or let us do it). I would > choose CMake, with the configure options: > > WITH_PYTHON=ON > PYTHON_EXECUTABLE=enthought python executable > PYTHON_INCLUDE_DIR=enthought python include path > PYTHON_LIBRARY=enthought python library > CMAKE_OSX_ARCHITECTURES= same as `file enthought python library` > > > I'm not in front of a mac today, but could produce some binaries for you > tomorrow, which version of EPD are you using? > > Thanks > Best > Frank > > > -----Original Message----- > > From: <a href="mailto:sbml-interoperability-bounces@caltech.edu" target="_blank">sbml-interoperability-bounces@caltech.edu</a> [mailto:sbml- > > <a href="mailto:interoperability-bounces@caltech.edu]" target="_blank">interoperability-bounces@caltech.edu]</a> On Behalf Of Johann Rohwer > > Sent: Wednesday, April 24, 2013 3:37 PM > > To: <a href="mailto:sbml-interoperability@caltech.edu" target="_blank">sbml-interoperability@caltech.edu</a> > > Subject: [sbml-interoperability] Installing libsbml on MacOSX > > > > I've been trying to install libsbml on MacOSX using the provided stable > > 5.8.0 > > > binary package installers. I need the python bindings of libsbml. > > Importing > > > libsbml from a system python prompt works fine. However, because we are > > interfacing with other python scientific computing software, we are using > > the > > > Entought Python Distribution (EPD) on these boxes. This is therefore a > > different python tree from the system python one. I can modify the python > > path so that the libsbml library files are found from within the EPD > > python > > > tree, but when I try to import libsbml from the EPD python prompt, it > > crashes > > > with a huge stack dump. > > > > Are the precompiled binaries of libsbml only for the system python on OSX? > > Does this mean that I have to recompile from sources and link to the EPD > > python at compile time using the > --with-python and --with-python- > > interpreter flags? > > > > I should add that I am a relative newbie to MacOSX related things but > > completely at home compiling things in Linux (the Macs are used in a > > teaching lab that we are setting up). So the concept of different python > > frameworks on the same machine is a bit foreign to me. > > > > Any pointers or ideas would be appreciated... > > > > Johann > > > > > > -- > > Johann M. Rohwer > > Professor of Systems Biology > > Dept. of Biochemistry, Stellenbosch University Private Bag X1, 7602 > > Matieland SOUTH AFRICA > > Tel +27-21-8085843 Fax +27-21-8085863 > > <a href="mailto:jr@sun.ac.za" target="_blank">jr@sun.ac.za</a> <a href="http://www.jjj.sun.ac.za" target="_blank">http://www.jjj.sun.ac.za</a> > > -- > > Registered Linux User #434886 > > -- > > E-pos vrywaringsklousule > > > > Hierdie e-pos mag vertroulike inligting bevat en mag regtens > > geprivilegeerd > > > wees en is slegs bedoel vir die persoon aan wie dit geadresseer is. Indien > > u > > > nie die bedoelde ontvanger is nie, word u hiermee in kennis gestel dat u > > hierdie dokument geensins mag gebruik, versprei of kopieer nie. Stel ook > > asseblief die sender onmiddellik per telefoon in kennis en vee die e-pos > > uit. > > > Die Universiteit aanvaar nie aanspreeklikheid vir enige skade, verlies of > > uitgawe wat voortspruit uit hierdie e-pos en/of die oopmaak van enige lC*s > > aangeheg by hierdie e-pos nie. > > > > E-mail disclaimer > > > > This e-mail may contain confidential information and may be legally > > privileged > > > and is intended only for the person to whom it is addressed. If you are > > not > > > the intended recipient, you are notified that you may not use, distribute > > or > > > copy this document in any manner whatsoever. Kindly also notify the sender > > immediately by telephone, and delete the e-mail. The University does not > > accept liability for any damage, loss or expense arising from this e-mail > > and/or accessing any files attached to this e-mail. > > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> -- Johann M. Rohwer Professor of Systems Biology Dept. of Biochemistry, Stellenbosch University Private Bag X1, 7602 Matieland SOUTH AFRICA Tel +27-21-8085843 Fax +27-21-8085863 <a href="mailto:jr@sun.ac.za" target="_blank">jr@sun.ac.za</a> <a href="http://www.jjj.sun.ac.za" target="_blank">http://www.jjj.sun.ac.za</a> -- Registered Linux User #434886 -- E-pos vrywaringsklousule Hierdie e-pos mag vertroulike inligting bevat en mag regtens geprivilegeerd wees en is slegs bedoel vir die persoon aan wie dit geadresseer is. Indien u nie die bedoelde ontvanger is nie, word u hiermee in kennis gestel dat u hierdie dokument geensins mag gebruik, versprei of kopieer nie. Stel ook asseblief die sender onmiddellik per telefoon in kennis en vee die e-pos uit. Die Universiteit aanvaar nie aanspreeklikheid vir enige skade, verlies of uitgawe wat voortspruit uit hierdie e-pos en/of die oopmaak van enige lÃªs aangeheg by hierdie e-pos nie. E-mail disclaimer This e-mail may contain confidential information and may be legally privileged and is intended only for the person to whom it is addressed. If you are not the intended recipient, you are notified that you may not use, distribute or copy this document in any manner whatsoever. Kindly also notify the sender immediately by telephone, and delete the e-mail. The University does not accept liability for any damage, loss or expense arising from this e-mail and/or accessing any files attached to this e-mail.

Re: Installing libsbml on MacOSX

fbergman — 2013-04-24T14:02:59-00:00

Hello Johann, Enthought does indeed use a different python than the one of the operating system. Since python bindings are not compatible among the different versions the crash is expected. So yes, you would have to compile them yourself (or let us do it). I would choose CMake, with the configure options: WITH_PYTHON=ON PYTHON_EXECUTABLE=enthought python executable PYTHON_INCLUDE_DIR=enthought python include path PYTHON_LIBRARY=enthought python library CMAKE_OSX_ARCHITECTURES= same as `file enthought python library` I'm not in front of a mac today, but could produce some binaries for you tomorrow, which version of EPD are you using? Thanks Best Frank > -----Original Message----- > From: <a href="mailto:sbml-interoperability-bounces@caltech.edu" target="_blank">sbml-interoperability-bounces@caltech.edu</a> [mailto:sbml- > <a href="mailto:interoperability-bounces@caltech.edu]" target="_blank">interoperability-bounces@caltech.edu]</a> On Behalf Of Johann Rohwer > Sent: Wednesday, April 24, 2013 3:37 PM > To: <a href="mailto:sbml-interoperability@caltech.edu" target="_blank">sbml-interoperability@caltech.edu</a> > Subject: [sbml-interoperability] Installing libsbml on MacOSX > > I've been trying to install libsbml on MacOSX using the provided stable 5.8.0 > binary package installers. I need the python bindings of libsbml. Importing > libsbml from a system python prompt works fine. However, because we are > interfacing with other python scientific computing software, we are using the > Entought Python Distribution (EPD) on these boxes. This is therefore a > different python tree from the system python one. I can modify the python > path so that the libsbml library files are found from within the EPD python > tree, but when I try to import libsbml from the EPD python prompt, it crashes > with a huge stack dump. > > Are the precompiled binaries of libsbml only for the system python on OSX? > Does this mean that I have to recompile from sources and link to the EPD > python at compile time using the > --with-python and --with-python- > interpreter flags? > > I should add that I am a relative newbie to MacOSX related things but > completely at home compiling things in Linux (the Macs are used in a > teaching lab that we are setting up). So the concept of different python > frameworks on the same machine is a bit foreign to me. > > Any pointers or ideas would be appreciated... > > Johann > > > -- > Johann M. Rohwer > Professor of Systems Biology > Dept. of Biochemistry, Stellenbosch University Private Bag X1, 7602 > Matieland SOUTH AFRICA > Tel +27-21-8085843 Fax +27-21-8085863 > <a href="mailto:jr@sun.ac.za" target="_blank">jr@sun.ac.za</a> <a href="http://www.jjj.sun.ac.za" target="_blank">http://www.jjj.sun.ac.za</a> > -- > Registered Linux User #434886 > -- > E-pos vrywaringsklousule > > Hierdie e-pos mag vertroulike inligting bevat en mag regtens geprivilegeerd > wees en is slegs bedoel vir die persoon aan wie dit geadresseer is. Indien u > nie die bedoelde ontvanger is nie, word u hiermee in kennis gestel dat u > hierdie dokument geensins mag gebruik, versprei of kopieer nie. Stel ook > asseblief die sender onmiddellik per telefoon in kennis en vee die e-pos uit. > Die Universiteit aanvaar nie aanspreeklikheid vir enige skade, verlies of > uitgawe wat voortspruit uit hierdie e-pos en/of die oopmaak van enige lC*s > aangeheg by hierdie e-pos nie. > > E-mail disclaimer > > This e-mail may contain confidential information and may be legally privileged > and is intended only for the person to whom it is addressed. If you are not > the intended recipient, you are notified that you may not use, distribute or > copy this document in any manner whatsoever. Kindly also notify the sender > immediately by telephone, and delete the e-mail. The University does not > accept liability for any damage, loss or expense arising from this e-mail > and/or accessing any files attached to this e-mail. ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Installing libsbml on MacOSX

jrohwer — 2013-04-24T13:37:01-00:00

I've been trying to install libsbml on MacOSX using the provided stable 5.8.0 binary package installers. I need the python bindings of libsbml. Importing libsbml from a system python prompt works fine. However, because we are interfacing with other python scientific computing software, we are using the Entought Python Distribution (EPD) on these boxes. This is therefore a different python tree from the system python one. I can modify the python path so that the libsbml library files are found from within the EPD python tree, but when I try to import libsbml from the EPD python prompt, it crashes with a huge stack dump. Are the precompiled binaries of libsbml only for the system python on OSX? Does this mean that I have to recompile from sources and link to the EPD python at compile time using the > --with-python and --with-python- interpreter flags? I should add that I am a relative newbie to MacOSX related things but completely at home compiling things in Linux (the Macs are used in a teaching lab that we are setting up). So the concept of different python frameworks on the same machine is a bit foreign to me. Any pointers or ideas would be appreciated... Johann -- Johann M. Rohwer Professor of Systems Biology Dept. of Biochemistry, Stellenbosch University Private Bag X1, 7602 Matieland SOUTH AFRICA Tel +27-21-8085843 Fax +27-21-8085863 <a href="mailto:jr@sun.ac.za" target="_blank">jr@sun.ac.za</a> <a href="http://www.jjj.sun.ac.za" target="_blank">http://www.jjj.sun.ac.za</a> -- Registered Linux User #434886 -- E-pos vrywaringsklousule Hierdie e-pos mag vertroulike inligting bevat en mag regtens geprivilegeerd wees en is slegs bedoel vir die persoon aan wie dit geadresseer is. Indien u nie die bedoelde ontvanger is nie, word u hiermee in kennis gestel dat u hierdie dokument geensins mag gebruik, versprei of kopieer nie. Stel ook asseblief die sender onmiddellik per telefoon in kennis en vee die e-pos uit. Die Universiteit aanvaar nie aanspreeklikheid vir enige skade, verlies of uitgawe wat voortspruit uit hierdie e-pos en/of die oopmaak van enige lÃªs aangeheg by hierdie e-pos nie. E-mail disclaimer This e-mail may contain confidential information and may be legally privileged and is intended only for the person to whom it is addressed. If you are not the intended recipient, you are notified that you may not use, distribute or copy this document in any manner whatsoever. Kindly also notify the sender immediately by telephone, and delete the e-mail. The University does not accept liability for any damage, loss or expense arising from this e-mail and/or accessing any files attached to this e-mail.

Re: SBML Toolbox, fbc, and groups

Ben Heavner — 2013-02-26T23:47:41-00:00

Thanks all, Sarah pointed me to the experimental branch of libsbml, which enables fbc support with the SBML Toolbox. So, I'm up and running (or at least stumbling and learning). Cheers! Ben

Re: SBML Toolbox, fbc, and groups

b.g.olivier — 2013-02-26T08:18:03-00:00

On 2013/02/25 11:58 PM, Lucian Smith wrote: > There are FBC and Comp models in the SBML Test suite now, so that would > probably work great: > > <a href="http://sbml.org/Software/SBML_Test_Suite" target="_blank">http://sbml.org/Software/SBML_Test_Suite</a> > > The FBC models are 1186-1196; the comp models are 1124-1183. > > Hi Ben Let me know if you'd like me to send you the test-suite model. You may also be interested in trying some models (most based on BiGG database models) that I've made available in the SBML3FBC and COBRA as part of a papers supplementary material. They can be found in the "Multi-format archives" section here: <a href="http://code.google.com/p/memesa-tools/" target="_blank">http://code.google.com/p/memesa-tools/</a> otherwise the test-suite model Cheers Brett > * Ben Heavner <<a href="mailto:bheavner@gmail.com" target="_blank">bheavner@gmail.com</a>> [2013-02-25 22:35] writes: >> Hi Sarah - >> >> Is there an easy way to check if the version of libsbml I'm using was built with fbc enabled (I downloaded a windows binary from sbml.org)? For example, is there a known-working sbml file that uses the fbc package that I can use to test my libsbml? >> >> As always, thanks for your efforts and great work! >> >> Ben >> >> Sarah Keating wrote on 23 Feb '13 08:28 >>> Hi Ben >>> >>> SBMLToolbox 4.1.0 does indeed support fbc. >>> >>> If you scroll to the bottom of >>> >>> <a href="http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_API_Manual" target="_blank">http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_API_Manual</a> >>> >>> you will see the additional functions and indeed how the structure has >>> been extended to include fbc concepts. >>> >>> This is the same structure that the matlab binding for libsbml uses when >>> built with fbc enabled. >>> >>> It does not as yet support groups and I'm not sure when I will have time >>> to add that support. It is also an issue with the libsbml matlab >>> bindings which are a little more complex to extend for packages that the >>> other language bindings; so we have not yet decided whether/how we will >>> do that. >>> >>> Sarah >>> >>> >>> >>> On 22/02/2013 22:59, Ben Heavner wrote: >>>> Hello all, >>>> >>>> Does the SBML Toolbox support the groups and fbc packages, or are there plans for such support? >>>> >>>> Specifically, how will the SBML Toolbox MATLAB data structure (<a href="http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_MATLAB_Structure" target="_blank">http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_MATLAB_Structure</a>) incorporate information from these packages? Or does it already? >>>> >>>> Thanks! >>>> Ben >>>> ____________________________________________________________ >>>> To manage your sbml-interoperability list subscription, visit >>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >>>> >>>> For a web interface to the sbml-interoperability mailing list, visit >>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>> >>>> For questions or feedback about the sbml-interoperability list, >>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>> >>>> >>> ____________________________________________________________ >>> To manage your sbml-interoperability list subscription, visit >>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >>> >>> For a web interface to the sbml-interoperability mailing list, visit >>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>> >>> For questions or feedback about the sbml-interoperability list, >>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >> >> ____________________________________________________________ >> To manage your sbml-interoperability list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >> >> For a web interface to the sbml-interoperability mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-interoperability list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> -- Dr Brett G. Olivier Researcher in Computational Systems Biology Dept. Molecular Cell Physiology VU University Amsterdam Email: b.g.olivier[AT]vu[dot]nl Tel: +31 (0)20 592 4265 ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: SBML Toolbox, fbc, and groups

Lucian Smith — 2013-02-25T22:58:58-00:00

There are FBC and Comp models in the SBML Test suite now, so that would probably work great: <a href="http://sbml.org/Software/SBML_Test_Suite" target="_blank">http://sbml.org/Software/SBML_Test_Suite</a> The FBC models are 1186-1196; the comp models are 1124-1183. -Lucian * Ben Heavner <<a href="mailto:bheavner@gmail.com" target="_blank">bheavner@gmail.com</a>> [2013-02-25 22:35] writes: > > Hi Sarah - > > Is there an easy way to check if the version of libsbml I'm using was built with fbc enabled (I downloaded a windows binary from sbml.org)? For example, is there a known-working sbml file that uses the fbc package that I can use to test my libsbml? > > As always, thanks for your efforts and great work! > > Ben > > Sarah Keating wrote on 23 Feb '13 08:28 > > Hi Ben > > > > SBMLToolbox 4.1.0 does indeed support fbc. > > > > If you scroll to the bottom of > > > > <a href="http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_API_Manual" target="_blank">http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_API_Manual</a> > > > > you will see the additional functions and indeed how the structure has > > been extended to include fbc concepts. > > > > This is the same structure that the matlab binding for libsbml uses when > > built with fbc enabled. > > > > It does not as yet support groups and I'm not sure when I will have time > > to add that support. It is also an issue with the libsbml matlab > > bindings which are a little more complex to extend for packages that the > > other language bindings; so we have not yet decided whether/how we will > > do that. > > > > Sarah > > > > > > > > On 22/02/2013 22:59, Ben Heavner wrote: > > > > > > Hello all, > > > > > > Does the SBML Toolbox support the groups and fbc packages, or are there plans for such support? > > > > > > Specifically, how will the SBML Toolbox MATLAB data structure (<a href="http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_MATLAB_Structure" target="_blank">http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_MATLAB_Structure</a>) incorporate information from these packages? Or does it already? > > > > > > Thanks! > > > Ben > > > ____________________________________________________________ > > > To manage your sbml-interoperability list subscription, visit > > > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > > > > > For a web interface to the sbml-interoperability mailing list, visit > > > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > > > > > For questions or feedback about the sbml-interoperability list, > > > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > > > > > > > > > ____________________________________________________________ > > To manage your sbml-interoperability list subscription, visit > > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > > > For a web interface to the sbml-interoperability mailing list, visit > > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > > > For questions or feedback about the sbml-interoperability list, > > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: SBML Toolbox, fbc, and groups

Ben Heavner — 2013-02-25T17:24:02-00:00

Hi Sarah - Is there an easy way to check if the version of libsbml I'm using was built with fbc enabled (I downloaded a windows binary from sbml.org)? For example, is there a known-working sbml file that uses the fbc package that I can use to test my libsbml? As always, thanks for your efforts and great work! Ben <table border="0" align="center" width="90%" cellpadding="3" cellspacing="1"><tr><td class="SmallText">Sarah Keating wrote on 23 Feb '13 08:28</td></tr><tr><td class="quote"> Hi Ben SBMLToolbox 4.1.0 does indeed support fbc. If you scroll to the bottom of <a href="http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_API_Manual" target="_blank">http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_API_Manual</a> you will see the additional functions and indeed how the structure has been extended to include fbc concepts. This is the same structure that the matlab binding for libsbml uses when built with fbc enabled. It does not as yet support groups and I'm not sure when I will have time to add that support. It is also an issue with the libsbml matlab bindings which are a little more complex to extend for packages that the other language bindings; so we have not yet decided whether/how we will do that. Sarah On 22/02/2013 22:59, Ben Heavner wrote: > > Hello all, > > Does the SBML Toolbox support the groups and fbc packages, or are there plans for such support? > > Specifically, how will the SBML Toolbox MATLAB data structure (<a href="http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_MATLAB_Structure" target="_blank">http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_MATLAB_Structure</a>) incorporate information from these packages? Or does it already? > > Thanks! > Ben > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> </td></tr></table>

Re: SBML Toolbox, fbc, and groups

Sarah Keating — 2013-02-23T13:28:23-00:00

Hi Ben SBMLToolbox 4.1.0 does indeed support fbc. If you scroll to the bottom of <a href="http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_API_Manual" target="_blank">http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_API_Manual</a> you will see the additional functions and indeed how the structure has been extended to include fbc concepts. This is the same structure that the matlab binding for libsbml uses when built with fbc enabled. It does not as yet support groups and I'm not sure when I will have time to add that support. It is also an issue with the libsbml matlab bindings which are a little more complex to extend for packages that the other language bindings; so we have not yet decided whether/how we will do that. Sarah On 22/02/2013 22:59, Ben Heavner wrote: > > Hello all, > > Does the SBML Toolbox support the groups and fbc packages, or are there plans for such support? > > Specifically, how will the SBML Toolbox MATLAB data structure (<a href="http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_MATLAB_Structure" target="_blank">http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_MATLAB_Structure</a>) incorporate information from these packages? Or does it already? > > Thanks! > Ben > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

SBML Toolbox, fbc, and groups

Ben Heavner — 2013-02-22T22:56:32-00:00

Hello all, Does the SBML Toolbox support the groups and fbc packages, or are there plans for such support? Specifically, how will the SBML Toolbox MATLAB data structure (<a href="http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_MATLAB_Structure" target="_blank">http://sbml.org/Software/SBMLToolbox/SBMLToolbox_4.0_MATLAB_Structure</a>) incorporate information from these packages? Or does it already? Thanks! Ben

Re: jSbml cannot parse an xml file.

mihajlv — 2013-02-14T14:32:45-00:00

Thank you both. I should have thought of using the validator first. Vidoje

Re: [sbml-interoperability] jSbml cannot parse an xml file.

Andreas Dräger — 2013-02-14T07:14:43-00:00

Dear Vidoje, I would just like to point out that there is a specialized list for questions regarding JSBML (<a href="mailto:jsbml-development@caltech.edu" target="_blank">jsbml-development@caltech.edu</a>; CC). Please e-mail questions to this library directly to this list. Thanks to Frank for the quick answer; in this particular case, the SBML interoperability mailing list was actually a good choice because the problem was rather the SBML code itself. If you have any further questions regarding JSBML, please don't hesitate to send an e-mail to <a href="mailto:jsbml-development@caltech.edu." target="_blank">jsbml-development@caltech.edu.</a> Cheers Andreas -- Dr. Andreas Dräger University of Tuebingen Center for Bioinformatics Tuebingen (ZBIT) Sand 1 72076 Tübingen Germany Phone: +49-7071-29-78982 Fax: +49-7071-29-5091 ____________________________________________________________ To manage your jsbml-development list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/jsbml-development" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/jsbml-development</a> For a web interface to the jsbml-development mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the jsbml-development list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: jSbml cannot parse an xml file.

Andreas Dräger — 2013-02-14T07:14:43-00:00

Dear Vidoje, I would just like to point out that there is a specialized list for questions regarding JSBML (<a href="mailto:jsbml-development@caltech.edu" target="_blank">jsbml-development@caltech.edu</a>; CC). Please e-mail questions to this library directly to this list. Thanks to Frank for the quick answer; in this particular case, the SBML interoperability mailing list was actually a good choice because the problem was rather the SBML code itself. If you have any further questions regarding JSBML, please don't hesitate to send an e-mail to <a href="mailto:jsbml-development@caltech.edu." target="_blank">jsbml-development@caltech.edu.</a> Cheers Andreas -- Dr. Andreas Dräger University of Tuebingen Center for Bioinformatics Tuebingen (ZBIT) Sand 1 72076 Tübingen Germany Phone: +49-7071-29-78982 Fax: +49-7071-29-5091 ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: jSbml cannot parse an xml file.

fbergman — 2013-02-14T06:49:27-00:00

Hello, I'm afraid the problem is not with JSBML, but rather with the model, when you use the online validator: <a href="http://sbml.org/Facilities/Validator/" target="_blank">http://sbml.org/Facilities/Validator/</a> you see that there are many things wrong with the model. To list just a view: - case is important in element / attribute names, the model violates this: + speciesreference -> speciesReference + initialamount -> initialAmount + listofcompartments -> listOfCompartments + listofspecies -> listOfSpecies + (same for reactions / reactants / products / modifiers) - identifiers contain invalid characters (all the 'name' elements in SBML L1V1 have to be identifiers containing no dashes, commas or brackets) - species attributre on speciesReference is invalid in SBML L1V1 If you like to take this offline, I'm sure we can get this to work. However, this will involve more changes. All the best Frank > -----Original Message----- > From: <a href="mailto:sbml-interoperability-bounces@caltech.edu" target="_blank">sbml-interoperability-bounces@caltech.edu</a> [mailto:sbml- > <a href="mailto:interoperability-bounces@caltech.edu]" target="_blank">interoperability-bounces@caltech.edu]</a> On Behalf Of mihajlv > Sent: Wednesday, February 13, 2013 10:04 PM > To: <a href="mailto:sbml-interoperability@caltech.edu" target="_blank">sbml-interoperability@caltech.edu</a> > Subject: [sbml-interoperability] jSbml cannot parse an xml file. > > Hello, > I have the latest jSBML installed and I am trying to parse the attached file but I > keep getting the many similar warnings: > > WARN (SBMLCoreParser.java:806) - The element compartment is not > recognized WARN (SBMLCoreParser.java:355) - Could not recognize the > attribute 'volume' on the element compartment. Please check the SBML > specifications. > WARN (SBMLCoreParser.java:806) - The element specie is not recognized ... > I've attached the document for viewing. If I use another file which is a higher > version it works, does it have to do anything with the version and how can I fix > this? > > Thank you for your time, > Vidoje ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

jSbml cannot parse an xml file.

mihajlv — 2013-02-13T20:27:39-00:00

Hello, I have the latest jSBML installed and I am trying to parse the attached file but I keep getting the many similar warnings: WARN (SBMLCoreParser.java:806) - The element compartment is not recognized WARN (SBMLCoreParser.java:355) - Could not recognize the attribute 'volume' on the element compartment. Please check the SBML specifications. WARN (SBMLCoreParser.java:806) - The element specie is not recognized ... I've attached the document for viewing. If I use another file which is a higher version it works, does it have to do anything with the version and how can I fix this? Thank you for your time, Vidoje

Re: libSBML MATLAB bindings

Sarah Keating — 2013-02-12T11:19:31-00:00

Hi Sebastian Thanks for your interest in SBML and sorry your are having difficulties. Unfortunately MATLAB and different Linux builds are a common cause of issues that arise and since we cannot maintain test beds for all versions we cannot actually replicate your problem. However, from your log, it does look like the libsbml that the MATLAB mex compilers is trying to link to is not the 32 bit library it was expecting. Can I suggest you try the following: 1. Download the 32-bit Linux rpm from the libsbml download site <a href="http://sourceforge.net/projects/sbml/files/libsbml/5.7.0/stable/Linux/32-bit/" target="_blank">http://sourceforge.net/projects/sbml/files/libsbml/5.7.0/stable/Linux/32-bit/</a> This will allow you to install a 32 bit binary for libSBML. This will have all the 32 bit dependencies that may not have been available on your 64 bit Xubuntu. 2. Open MATLAB Navigate to the libsbml/src/bindings/matlab directory you already have. Within MATLAB run buildSBML('fullpath to libsbml include directory', 'fullpath to the location of the library files') This will ensure the build picks up the 32 bit libraries you installed from the rpm ============== Hopefully this will work for you. Please let us know if it does not and we can try and think of another solution. Thanks Sarah (SBML Team) ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

libSBML MATLAB bindings

Sebastian Jaramillo Riveri — 2013-02-10T20:48:35-00:00

Dear all, I am trying to build the bindings to libSBML from matlab, however I am getting an error and I am not quite sure how to proceed. I was hoping some of you could guide me. Any orientation will be kindly welcomed. I have Xubuntu 64bit, however my version of MATLAB is 32bit (so I used the flag --enable-m32). I tried first compiling with the --with-matlab flag, however that was not possible: Failed to determine the mex extension, that means that Matlab could not be started. This is possibly due to license issues (or a missing VPN connection.) Please start Matlab manually and run buildSBML.m (in src/bindings/matlab). Now, here is the output from evaluating buildSBML and installSBML: EDU>> buildSBML * Doing preliminary checks of runtime environment ... - This appears to be MATLAB and not Octave. - MATLAB reports the OS is 32-bit Linux. Constructing the libSBML MATLAB interface. * Trying to establish our location ... - We are in the libSBML subdirectory for Matlab. - We appear to be in the libSBML source tree. - We have write access here! That makes us happy. * Locating libSBML library and include files ... - Found /home/seba/code/sbml/libSBML-5.7.0-Source/src/.libs/libsbml.so - Root of includes is /home/seba/code/sbml/libSBML-5.7.0-Source/src * Creating mex files in /home/seba/code/sbml/libSBML-5.7.0-Source/src/bindings/matlab - Building TranslateSBML ... Warning: You are using gcc version "4.6.1-9ubuntu3)". The version currently supported with MEX is "4.3.4". For a list of currently supported compilers see: <a href="http://www.mathworks.com/support/compilers/current_release/" target="_blank">http://www.mathworks.com/support/compilers/current_release/</a> - Building OutputSBML ... Warning: You are using gcc version "4.6.1-9ubuntu3)". The version currently supported with MEX is "4.3.4". For a list of currently supported compilers see: <a href="http://www.mathworks.com/support/compilers/current_release/" target="_blank">http://www.mathworks.com/support/compilers/current_release/</a> OutputSBML.c: In function â€˜mexFunctionâ€™: OutputSBML.c:193:14: warning: assignment discards â€˜constâ€™ qualifier from pointer target type [enabled by default] Successfully finished. If appropriate, please run "installSBML" next. EDU>> installSBML Installing the libSBML interface. * Doing preliminary checks of runtime environment ... - This appears to be MATLAB and not Octave. Checking for executables in /usr/local/lib Executables found Adding /usr/local/lib to path ... Success Attempting to execute functions TranslateSBML failed Invalid MEX-file '/usr/local/lib/TranslateSBML.mexglx': libsbml.so.5: wrong ELF class: ELFCLASS64 Removing /usr/local/lib from path ... Success ??? Error using ==> installSBML at 72 Executables were located at /usr/local/lib but failed to execute Please contact the libSBML team <a href="mailto:libsbml-team@caltech.edu" target="_blank">libsbml-team@caltech.edu</a> for further assistance Thank you in advance for you help, All the best. -- Sebastian. ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

The SBML Test Suite Database is now available

Mike Hucka — 2013-01-16T22:59:10-00:00

Ladies and gentlemen, The SBML Team is pleased to announce that the SBML Test Suite Database is now online. The SBML Test Suite Database publicizes software systems' support of SBML as determined by their performance on the SBML Test Suite. As a user of SBML and SBML software, you can use this SBML Test Suite Database to find the test results for different software systems. If you are a developer of SBML Software you can also submit results to this database and thereby make them known to the world. The front page is at <a href="http://sbml.org/Facilities/Database/" target="_blank">http://sbml.org/Facilities/Database/</a> We are opening the database now with one set of example results, but many developers have started submitting their own, so by the time you read this, several more tools may already be represented in the database. Please report any issues or questions you have and we will work to improve the system. We want to make this system as useful for the SBML community as possible. Credit for the bulk of the work on this system goes to Frank Bergmann. Best regards, The SBML Team ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: package removal with libSBML

fbergman — 2013-01-03T17:00:25-00:00

> I've been trying to use the libSBML SBMLStripPackageConverter to remove > package constructs from L3 SBML documents, but without success. Does > anyone have example code I could plagiarise?- preferably (but not > necessarily) in python. > indeed there is, as part of the *upcoming* libsbml release we do have examples for the conversion API for all binding languages. It is already in SVN as stripPackage example: <a href="http://sbml.svn.sourceforge.net/viewvc/sbml/trunk/libsbml/examples/" target="_blank">http://sbml.svn.sourceforge.net/viewvc/sbml/trunk/libsbml/examples/</a> for example in python: <a href="http://sbml.svn.sourceforge.net/viewvc/sbml/trunk/libsbml/examples/python/stripPackage.py?revision=16719&view=markup" target="_blank">http://sbml.svn.sourceforge.net/viewvc/sbml/trunk/libsbml/examples/python/stripPackage.py?revision=16719&view=markup</a> > On a related note, I see there is a SBMLLevelVersionConverter. Does this > offer any functionality above setLevelAndVersion? > when calling setLevelAndVersion (a function that has been there for a long time), it is actually setting up a converter in the background that will be executed. The only advantage of running it yourself (such as done in the layout / convertLayout example) is that you can pass an additional option 'ignore packages' that will allow the conversion to finish despite some packages being present. (Willingly acknowledging, that that will mean that all information in those packages will be (most likely) lost). best Frank > As ever, thanks for the help, > k > > -- > kieran smallbone > <a href="mailto:kieran.smallbone@manchester.ac.uk" target="_blank">kieran.smallbone@manchester.ac.uk</a> > <a href="http://u003f.com/" target="_blank">http://u003f.com/</a> > > manchester centre for integrative systems biology > manchester interdisciplinary biocentre > 131 princess street > manchester m1 7dn. uk > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

package removal with libSBML

kieransmallbone — 2013-01-03T16:49:00-00:00

Hello folks, I've been trying to use the libSBML SBMLStripPackageConverter to remove package constructs from L3 SBML documents, but without success. Does anyone have example code I could plagiarise?- preferably (but not necessarily) in python. On a related note, I see there is a SBMLLevelVersionConverter. Does this offer any functionality above setLevelAndVersion? As ever, thanks for the help, k -- kieran smallbone <a href="mailto:kieran.smallbone@manchester.ac.uk" target="_blank">kieran.smallbone@manchester.ac.uk</a> <a href="http://u003f.com/" target="_blank">http://u003f.com/</a> manchester centre for integrative systems biology manchester interdisciplinary biocentre 131 princess street manchester m1 7dn. uk ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: package installation

fbergman — 2013-01-02T13:49:03-00:00

> Thanks for the quick reply Frank. I figured out why I couldn't get the > binary to work: the default installation directory from python.org is > not where the binary expects. > > I got around this by adding a symbolic link > $ ln -s /Library/Frameworks/Python.framework/ > /System/Library/Frameworks/Python.framework > then everything was rosy. > I'm not sure that is save. For all binaries for libsbml we use the systems python installation, so if that is not the one to use you might be better of in compiling it yourself. The important variables to watch for will be PYTHON_EXECUTABLE, PYTHON_INCLUDE_DIR and PYTHON_LIBRARY, they have to be set to the correct libraries. > As an aside though, my first pass at configuring with cmake didn't work > out: there were lots of errors of the type > > libSBML_SOURCE/src/bindings/python/local-downcast-packages-layout.cpp:81: error: > ‘SWIGTYPE_p_GeneralGlyph’ was not declared in this scope > I'm not sure how this can come about. That would mean, that while configuring you had ENABLE_LAYOUT=ON, but during the subsequent compile run, layout was turned off again. Maybe we should take this offline and you could mail me your CMakeCache.txt (from your build dir) then i could see what was going on. thanks Frank > k > > On 02/01/2013 12:18, Frank T. Bergmann wrote: >>> Could you give me a duffers guide to installation of packages with >>> libSBML please? I downloaded the source from >>> <a href="http://sourceforge.net/projects/sbml/files/libsbml/5.6.0/experimental/2012-08-07/src/" target="_blank">http://sourceforge.net/projects/sbml/files/libsbml/5.6.0/experimental/2012-08-07/src/</a> >>> and tried >>> ./configure --with-python --with-matlab=/Applications/MATLAB_R2012a.app/ >>> but the python command >>>>> sbmlns = libsbml.SBMLNamespaces(3, 1, "groups", 1) >>> runs into >>>>> ValueError: groups : No such package registered. >>> >>> Am I missing a step? >>> >> >> I'm afraid you will need cmake to configure a libsbml build with packages or rely on the binary libsbml-5.6.0-Darwin.dmg that is in the binaries directory. >> >> For more information see: >> >> <a href="http://sbml.org/Software/libSBML/docs/cpp-api/libsbml-installation.html#unix-cmake" target="_blank">http://sbml.org/Software/libSBML/docs/cpp-api/libsbml-installation.html#unix-cmake</a> >> >> basically all you would need to do for your options above (assuming cmake is installed) is to create build directory, in which all binary files will be located: >> >> mkdir build >> cd build >> cmake -DWITH_PYTHON=ON -DWITH_MATLAB=ON -DMATLAB_ROOT_PATH=/Applications/MATLAB_R2012a.app/ >> >> make >> >> Please let me know if you have further questions. >> >> all the best >> Frank >> ____________________________________________________________ >> To manage your sbml-interoperability list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >> >> For a web interface to the sbml-interoperability mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-interoperability list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >> > > -- > kieran smallbone > <a href="mailto:kieran.smallbone@manchester.ac.uk" target="_blank">kieran.smallbone@manchester.ac.uk</a> > <a href="http://u003f.com/" target="_blank">http://u003f.com/</a> > > manchester centre for integrative systems biology > manchester interdisciplinary biocentre > 131 princess street > manchester m1 7dn. uk > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: package installation

kieransmallbone — 2013-01-02T13:22:31-00:00

Thanks for the quick reply Frank. I figured out why I couldn't get the binary to work: the default installation directory from python.org is not where the binary expects. I got around this by adding a symbolic link $ ln -s /Library/Frameworks/Python.framework/ /System/Library/Frameworks/Python.framework then everything was rosy. As an aside though, my first pass at configuring with cmake didn't work out: there were lots of errors of the type libSBML_SOURCE/src/bindings/python/local-downcast-packages-layout.cpp:81: error: ‘SWIGTYPE_p_GeneralGlyph’ was not declared in this scope k On 02/01/2013 12:18, Frank T. Bergmann wrote: >> Could you give me a duffers guide to installation of packages with >> libSBML please? I downloaded the source from >> <a href="http://sourceforge.net/projects/sbml/files/libsbml/5.6.0/experimental/2012-08-07/src/" target="_blank">http://sourceforge.net/projects/sbml/files/libsbml/5.6.0/experimental/2012-08-07/src/</a> >> and tried >> ./configure --with-python --with-matlab=/Applications/MATLAB_R2012a.app/ >> but the python command >>>> sbmlns = libsbml.SBMLNamespaces(3, 1, "groups", 1) >> runs into >>>> ValueError: groups : No such package registered. >> >> Am I missing a step? >> > > I'm afraid you will need cmake to configure a libsbml build with packages or rely on the binary libsbml-5.6.0-Darwin.dmg that is in the binaries directory. > > For more information see: > > <a href="http://sbml.org/Software/libSBML/docs/cpp-api/libsbml-installation.html#unix-cmake" target="_blank">http://sbml.org/Software/libSBML/docs/cpp-api/libsbml-installation.html#unix-cmake</a> > > basically all you would need to do for your options above (assuming cmake is installed) is to create build directory, in which all binary files will be located: > > mkdir build > cd build > cmake -DWITH_PYTHON=ON -DWITH_MATLAB=ON -DMATLAB_ROOT_PATH=/Applications/MATLAB_R2012a.app/ > > make > > Please let me know if you have further questions. > > all the best > Frank > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > -- kieran smallbone <a href="mailto:kieran.smallbone@manchester.ac.uk" target="_blank">kieran.smallbone@manchester.ac.uk</a> <a href="http://u003f.com/" target="_blank">http://u003f.com/</a> manchester centre for integrative systems biology manchester interdisciplinary biocentre 131 princess street manchester m1 7dn. uk ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: package installation

fbergman — 2013-01-02T12:18:00-00:00

> Could you give me a duffers guide to installation of packages with > libSBML please? I downloaded the source from > <a href="http://sourceforge.net/projects/sbml/files/libsbml/5.6.0/experimental/2012-08-07/src/" target="_blank">http://sourceforge.net/projects/sbml/files/libsbml/5.6.0/experimental/2012-08-07/src/</a> > and tried > ./configure --with-python --with-matlab=/Applications/MATLAB_R2012a.app/ > but the python command >>> sbmlns = libsbml.SBMLNamespaces(3, 1, "groups", 1) > runs into >>> ValueError: groups : No such package registered. > > Am I missing a step? > I'm afraid you will need cmake to configure a libsbml build with packages or rely on the binary libsbml-5.6.0-Darwin.dmg that is in the binaries directory. For more information see: <a href="http://sbml.org/Software/libSBML/docs/cpp-api/libsbml-installation.html#unix-cmake" target="_blank">http://sbml.org/Software/libSBML/docs/cpp-api/libsbml-installation.html#unix-cmake</a> basically all you would need to do for your options above (assuming cmake is installed) is to create build directory, in which all binary files will be located: mkdir build cd build cmake -DWITH_PYTHON=ON -DWITH_MATLAB=ON -DMATLAB_ROOT_PATH=/Applications/MATLAB_R2012a.app/ make Please let me know if you have further questions. all the best Frank ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

package installation

kieransmallbone — 2013-01-02T12:07:12-00:00

Hello folks, Could you give me a duffers guide to installation of packages with libSBML please? I downloaded the source from <a href="http://sourceforge.net/projects/sbml/files/libsbml/5.6.0/experimental/2012-08-07/src/" target="_blank">http://sourceforge.net/projects/sbml/files/libsbml/5.6.0/experimental/2012-08-07/src/</a> and tried ./configure --with-python --with-matlab=/Applications/MATLAB_R2012a.app/ but the python command >> sbmlns = libsbml.SBMLNamespaces(3, 1, "groups", 1) runs into >> ValueError: groups : No such package registered. Am I missing a step? Cheers, k PS - I seem to have broken the online validator, sorry ... -- kieran smallbone <a href="mailto:kieran.smallbone@manchester.ac.uk" target="_blank">kieran.smallbone@manchester.ac.uk</a> <a href="http://u003f.com/" target="_blank">http://u003f.com/</a> manchester centre for integrative systems biology manchester interdisciplinary biocentre 131 princess street manchester m1 7dn. uk ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Request for new windows libsbml install executable

Ben Heavner — 2012-12-06T17:01:40-00:00

Arg. RTFM is a fine response to my question. Sorry. " libSBML-5.6.0-win-x64.exe Complete libSBML binary distribution, containing the precompiled C/C++ libSBML core library and the following libSBML language interfaces: C# (Mono), Java, MATLAB, Perl, Python, Ruby. (Note: the Octave interface is not included because there are no 64-bit versions of Octave available at this time.)" Whoops! -b

Re: Request for new windows libsbml install executable

Ben Heavner — 2012-12-06T16:59:34-00:00

I'm confused, though - I want to use libsbml with python and with matlab. Currently, the following executables are available: libSBML-5.6.0-win-matlab-x64.exe (which I think has MATLAB bindings) and libSBML-5.6.0-win-py2.7-amd64.exe (which I think has python bindings) I'm hoping to find something like libSBML-5.6.0-win-matlab-py2.7-amd64.exe so that I can use both. Does such a thing exist, or is there a different way to use libSBML with both matlab and python? Thanks! Ben

Re: Request for new windows libsbml install executable

Ben Heavner — 2012-12-06T16:55:57-00:00

Thanks very much Frank!

Re: Request for new windows libsbml install executable

fbergman — 2012-12-06T08:16:40-00:00

Hello Ben, Here you go: <a href="http://sourceforge.net/projects/sbml/files/libsbml/5.6.0/stable/Windows/64-bit/" target="_blank">http://sourceforge.net/projects/sbml/files/libsbml/5.6.0/stable/Windows/64-bit/</a> if you also would like the SBML L3 packages enabled the link is: <a href="http://sourceforge.net/projects/sbml/files/libsbml/5.6.0/experimental/2012-08-07/binaries/" target="_blank">http://sourceforge.net/projects/sbml/files/libsbml/5.6.0/experimental/2012-08-07/binaries/</a> best Frank On Dec 6, 2012, at 2:05 AM, Ben Heavner <<a href="mailto:bheavner@gmail.com" target="_blank">bheavner@gmail.com</a>> wrote: > > So I'm following the herd and trying to learn a bit about python. However, I'm not quite ready to give up my currently-more-productive MATLAB. > > I'm using Windows 7 these days, and would prefer to avoid previous pain of trying to compile libSBML for 64-bit Windows using the free version of MSVC++. > > Could anyone provide an executable 64-bit windows libsbml 5.6.0 with MATLAB and python bindings enabled? > > Thanks in advance, > Ben > > p.s. I seem to have succeed at getting scipy and numpy working with 64-bit python 2.7.3 by using the unofficial versions from <a href="http://www.lfd.uci.edu/~gohlke/pythonlibs" target="_blank">http://www.lfd.uci.edu/~gohlke/pythonlibs</a> > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Request for new windows libsbml install executable

Ben Heavner — 2012-12-06T00:38:17-00:00

So I'm following the herd and trying to learn a bit about python. However, I'm not quite ready to give up my currently-more-productive MATLAB. I'm using Windows 7 these days, and would prefer to avoid previous pain of trying to compile libSBML for 64-bit Windows using the free version of MSVC++. Could anyone provide an executable 64-bit windows libsbml 5.6.0 with MATLAB and python bindings enabled? Thanks in advance, Ben p.s. I seem to have succeed at getting scipy and numpy working with 64-bit python 2.7.3 by using the unofficial versions from <a href="http://www.lfd.uci.edu/~gohlke/pythonlibs" target="_blank">http://www.lfd.uci.edu/~gohlke/pythonlibs</a>

Re: Issues installing libsbml and Python bindings in OS X and EPD

fbergman — 2012-11-23T19:44:11-00:00

> I'm pretty sure the problem at this point is an installation one - for example, there *is* no libsbml directory in site-packages, where it should be with make install. > indeed this will be created when running the setup.py script. > What's the best way to have someone go over the install logs? A large text-dump here, an attachment or something else? > the information we need at this point are: - config.log - please start IDLE and print the value of: import sys; sys.prefix Feel free to send it to me personally thanks Frank > Thanks, > > Eric > > On Nov 23, 2012, at 6:10 AM, Frank T. Bergmann wrote: > >> Sorry for the delay in answering this. With libsbml 5.6.0 we made sure the >> python bindings would work with EPD. Of course not having a license we could >> only test it with the free version (which was 32bit only). That version >> installed into: >> >> /Library/Frameworks/Python.framework/Versions/7.3 >> >> As can be found out by running IDLE and running sys.prefix. So the >> configure line I uses would be: >> >> ./configure --with-python=/Library/Frameworks/Python.framework/Versions/7.3 >> --with-check --enable-universal-binary="-arch i386" >> >> Here the first is the prefix given by sys.prefix. The second argument is >> used so that I can run the libsbml tests for the python bindings (requires >> libcheck). And the third argument is so that even though I'm running on a >> 64bit system only the 32bit binaries are created. >> >> After running make the binaries are created in src/bindings/python, and (if >> --with-check) was given, the tests run fine with EPD. >> >> Running 'make install' will locate the binaries (both libsbml and the python >> bindings)in /usr/local/lib/python2.7/site-packages/libsbml. This requires >> you to set the PYTHONPATH. If this is not what you want, you can choose to >> additinally install it right into EPD by running: >> >> /Library/Frameworks/Python.framework/Versions/7.3/bin/python setup.py >> install --skip-build >> >>> From the src/bindings/python directory (the core libsbml library will still >> need to be installed into /usr/local/lib). >> >> Please let me know in case this does not help, >> >> All the best >> Frank >> >> >>> -----Original Message----- >>> From: <a href="mailto:sbml-interoperability-bounces@caltech.edu" target="_blank">sbml-interoperability-bounces@caltech.edu</a> [mailto:sbml- >>> <a href="mailto:interoperability-bounces@caltech.edu]" target="_blank">interoperability-bounces@caltech.edu]</a> On Behalf Of Lofgren, Eric >>> Sent: Thursday, November 22, 2012 11:08 PM >>> To: SBML Software Interoperability Discussion List >>> Subject: [sbml-interoperability] Issues installing libsbml and Python >> bindings >>> in OS X and EPD >>> >>> This appears to have actually been an installation issue - searching for >>> libsmbl.py only brings up the copies sitting in the downloaded >> libsbml-5.6.0 >>> folder on my desktop, rather than where they should be in the OS X file >>> system. That suggests to me that something's not installing correctly. >>> >>> I've run .configure along with make and make install - what terminal >> output >>> would be useful in diagnosing an installation problem? >>> >>> Eric >>> >>> On Nov 22, 2012, at 5:58 AM, Nicolas Rodriguez wrote: >>> >>>> On 11/22/2012 10:51 AM, Nicolas Rodriguez wrote: >>>>> This is probably a PYTHONPATH problem, search for the location of >>>>> libsbml.py and add the folder to the PYTHONPATH. >>>>> Then python will try to load the libsbml libraries and you might have >>>>> further problems. >>>> >>>> more infos can be found there : >>>> >>>> <a href="http://sbml.org/Software/libSBML/docs/python-api/libsbml-accessing.htm" target="_blank">http://sbml.org/Software/libSBML/docs/python-api/libsbml-accessing.htm</a> >>>> l#accessing-python >>>> >>>> Be sure to set properly 'DYLD_LIBRARY_PATH' as well. >>>> >>>>> >>>>> Nico >>>>> >>>>> >>>>> On 11/22/2012 12:34 AM, Lofgren, Eric wrote: >>>>>> I'm attempting to install libsbml with the accompanying Python >> bindings, >>> and am having a fair amount of trouble - and to be honest, I'm not sure >> quite >>> where things are going wrong. I'm using the Enthought Python Distribution, >>> which does admittedly add a layer of complexity to things. >>>>>> >>>>>> Following some advice from their listserv, I set up the configuration >> of >>> libsbml 5.6.0 using the following: >>>>>> >>>>>> ./configure >>>>>> --prefix=/Library/Frameworks/EPD64.framework/Versions/Current >>>>>> --with-python=/Library/Frameworks/EPD64.framework/Versions/Current >>>>>> >>>>>> Followed with: >>>>>> make >>>>>> and >>>>>> make install >>>>>> >>>>>> After that, I attempted to use libsbml's python bindings with the >> usual >>> from libsbml import * and got the following error: >>>>>> >>>>>>>>> from libsbml import * >>>>>> Traceback (most recent call last): >>>>>> File "", line 1, in >>>>>> ImportError: No module named libsbml >>>>>> >>>>>> Either libsbml isn't being installed correctly, or Python/EPD is >> having issues >>> finding the module. I can post the output from make and make install, but >> I >>> figured before I hit folks with a wall of text, I'd see if anyone had a >> solution >>> they knew of. >>>>>> >>>>>> Thanks, >>>>>> >>>>>> Eric >>>>>> ____________________________________________________________ >>>>>> To manage your sbml-interoperability list subscription, visit >>>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >>>>>> >>>>>> For a web interface to the sbml-interoperability mailing list, visit >>>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>>> >>>>>> For questions or feedback about the sbml-interoperability list, >>>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>> ____________________________________________________________ >>>>> To manage your sbml-interoperability list subscription, visit >>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >>>>> >>>>> For a web interface to the sbml-interoperability mailing list, visit >>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>> >>>>> For questions or feedback about the sbml-interoperability list, >>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>> >>>> ____________________________________________________________ >>>> To manage your sbml-interoperability list subscription, visit >>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >>>> >>>> For a web interface to the sbml-interoperability mailing list, visit >>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>> >>>> For questions or feedback about the sbml-interoperability list, >>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>> >>> ____________________________________________________________ >>> To manage your sbml-interoperability list subscription, visit >>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >>> >>> For a web interface to the sbml-interoperability mailing list, visit >>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>> >>> For questions or feedback about the sbml-interoperability list, contact >> sbml- >>> <a href="mailto:team@caltech.edu" target="_blank">team@caltech.edu</a> >> >> ____________________________________________________________ >> To manage your sbml-interoperability list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >> >> For a web interface to the sbml-interoperability mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-interoperability list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Issues installing libsbml and Python bindings in OS X and EPD

elofgren — 2012-11-23T17:30:19-00:00

I'm pretty sure the problem at this point is an installation one - for example, there *is* no libsbml directory in site-packages, where it should be with make install. What's the best way to have someone go over the install logs? A large text-dump here, an attachment or something else? Thanks, Eric On Nov 23, 2012, at 6:10 AM, Frank T. Bergmann wrote: > Sorry for the delay in answering this. With libsbml 5.6.0 we made sure the > python bindings would work with EPD. Of course not having a license we could > only test it with the free version (which was 32bit only). That version > installed into: > > /Library/Frameworks/Python.framework/Versions/7.3 > > As can be found out by running IDLE and running sys.prefix. So the > configure line I uses would be: > > ./configure --with-python=/Library/Frameworks/Python.framework/Versions/7.3 > --with-check --enable-universal-binary="-arch i386" > > Here the first is the prefix given by sys.prefix. The second argument is > used so that I can run the libsbml tests for the python bindings (requires > libcheck). And the third argument is so that even though I'm running on a > 64bit system only the 32bit binaries are created. > > After running make the binaries are created in src/bindings/python, and (if > --with-check) was given, the tests run fine with EPD. > > Running 'make install' will locate the binaries (both libsbml and the python > bindings)in /usr/local/lib/python2.7/site-packages/libsbml. This requires > you to set the PYTHONPATH. If this is not what you want, you can choose to > additinally install it right into EPD by running: > > /Library/Frameworks/Python.framework/Versions/7.3/bin/python setup.py > install --skip-build > >> From the src/bindings/python directory (the core libsbml library will still > need to be installed into /usr/local/lib). > > Please let me know in case this does not help, > > All the best > Frank > > >> -----Original Message----- >> From: <a href="mailto:sbml-interoperability-bounces@caltech.edu" target="_blank">sbml-interoperability-bounces@caltech.edu</a> [mailto:sbml- >> <a href="mailto:interoperability-bounces@caltech.edu]" target="_blank">interoperability-bounces@caltech.edu]</a> On Behalf Of Lofgren, Eric >> Sent: Thursday, November 22, 2012 11:08 PM >> To: SBML Software Interoperability Discussion List >> Subject: [sbml-interoperability] Issues installing libsbml and Python > bindings >> in OS X and EPD >> >> This appears to have actually been an installation issue - searching for >> libsmbl.py only brings up the copies sitting in the downloaded > libsbml-5.6.0 >> folder on my desktop, rather than where they should be in the OS X file >> system. That suggests to me that something's not installing correctly. >> >> I've run .configure along with make and make install - what terminal > output >> would be useful in diagnosing an installation problem? >> >> Eric >> >> On Nov 22, 2012, at 5:58 AM, Nicolas Rodriguez wrote: >> >>> On 11/22/2012 10:51 AM, Nicolas Rodriguez wrote: >>>> This is probably a PYTHONPATH problem, search for the location of >>>> libsbml.py and add the folder to the PYTHONPATH. >>>> Then python will try to load the libsbml libraries and you might have >>>> further problems. >>> >>> more infos can be found there : >>> >>> <a href="http://sbml.org/Software/libSBML/docs/python-api/libsbml-accessing.htm" target="_blank">http://sbml.org/Software/libSBML/docs/python-api/libsbml-accessing.htm</a> >>> l#accessing-python >>> >>> Be sure to set properly 'DYLD_LIBRARY_PATH' as well. >>> >>>> >>>> Nico >>>> >>>> >>>> On 11/22/2012 12:34 AM, Lofgren, Eric wrote: >>>>> I'm attempting to install libsbml with the accompanying Python > bindings, >> and am having a fair amount of trouble - and to be honest, I'm not sure > quite >> where things are going wrong. I'm using the Enthought Python Distribution, >> which does admittedly add a layer of complexity to things. >>>>> >>>>> Following some advice from their listserv, I set up the configuration > of >> libsbml 5.6.0 using the following: >>>>> >>>>> ./configure >>>>> --prefix=/Library/Frameworks/EPD64.framework/Versions/Current >>>>> --with-python=/Library/Frameworks/EPD64.framework/Versions/Current >>>>> >>>>> Followed with: >>>>> make >>>>> and >>>>> make install >>>>> >>>>> After that, I attempted to use libsbml's python bindings with the > usual >> from libsbml import * and got the following error: >>>>> >>>>>>>> from libsbml import * >>>>> Traceback (most recent call last): >>>>> File "", line 1, in >>>>> ImportError: No module named libsbml >>>>> >>>>> Either libsbml isn't being installed correctly, or Python/EPD is > having issues >> finding the module. I can post the output from make and make install, but > I >> figured before I hit folks with a wall of text, I'd see if anyone had a > solution >> they knew of. >>>>> >>>>> Thanks, >>>>> >>>>> Eric >>>>> ____________________________________________________________ >>>>> To manage your sbml-interoperability list subscription, visit >>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >>>>> >>>>> For a web interface to the sbml-interoperability mailing list, visit >>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>> >>>>> For questions or feedback about the sbml-interoperability list, >>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>> ____________________________________________________________ >>>> To manage your sbml-interoperability list subscription, visit >>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >>>> >>>> For a web interface to the sbml-interoperability mailing list, visit >>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>> >>>> For questions or feedback about the sbml-interoperability list, >>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>> >>> ____________________________________________________________ >>> To manage your sbml-interoperability list subscription, visit >>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >>> >>> For a web interface to the sbml-interoperability mailing list, visit >>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>> >>> For questions or feedback about the sbml-interoperability list, >>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >> >> ____________________________________________________________ >> To manage your sbml-interoperability list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >> >> For a web interface to the sbml-interoperability mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-interoperability list, contact > sbml- >> <a href="mailto:team@caltech.edu" target="_blank">team@caltech.edu</a> > > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Issues installing libsbml and Python bindings in OS X and EPD

fbergman — 2012-11-23T11:10:02-00:00

Sorry for the delay in answering this. With libsbml 5.6.0 we made sure the python bindings would work with EPD. Of course not having a license we could only test it with the free version (which was 32bit only). That version installed into: /Library/Frameworks/Python.framework/Versions/7.3 As can be found out by running IDLE and running sys.prefix. So the configure line I uses would be: ./configure --with-python=/Library/Frameworks/Python.framework/Versions/7.3 --with-check --enable-universal-binary="-arch i386" Here the first is the prefix given by sys.prefix. The second argument is used so that I can run the libsbml tests for the python bindings (requires libcheck). And the third argument is so that even though I'm running on a 64bit system only the 32bit binaries are created. After running make the binaries are created in src/bindings/python, and (if --with-check) was given, the tests run fine with EPD. Running 'make install' will locate the binaries (both libsbml and the python bindings)in /usr/local/lib/python2.7/site-packages/libsbml. This requires you to set the PYTHONPATH. If this is not what you want, you can choose to additinally install it right into EPD by running: /Library/Frameworks/Python.framework/Versions/7.3/bin/python setup.py install --skip-build >From the src/bindings/python directory (the core libsbml library will still need to be installed into /usr/local/lib). Please let me know in case this does not help, All the best Frank > -----Original Message----- > From: <a href="mailto:sbml-interoperability-bounces@caltech.edu" target="_blank">sbml-interoperability-bounces@caltech.edu</a> [mailto:sbml- > <a href="mailto:interoperability-bounces@caltech.edu]" target="_blank">interoperability-bounces@caltech.edu]</a> On Behalf Of Lofgren, Eric > Sent: Thursday, November 22, 2012 11:08 PM > To: SBML Software Interoperability Discussion List > Subject: [sbml-interoperability] Issues installing libsbml and Python bindings > in OS X and EPD > > This appears to have actually been an installation issue - searching for > libsmbl.py only brings up the copies sitting in the downloaded libsbml-5.6.0 > folder on my desktop, rather than where they should be in the OS X file > system. That suggests to me that something's not installing correctly. > > I've run .configure along with make and make install - what terminal output > would be useful in diagnosing an installation problem? > > Eric > > On Nov 22, 2012, at 5:58 AM, Nicolas Rodriguez wrote: > > > On 11/22/2012 10:51 AM, Nicolas Rodriguez wrote: > >> This is probably a PYTHONPATH problem, search for the location of > >> libsbml.py and add the folder to the PYTHONPATH. > >> Then python will try to load the libsbml libraries and you might have > >> further problems. > > > > more infos can be found there : > > > > <a href="http://sbml.org/Software/libSBML/docs/python-api/libsbml-accessing.htm" target="_blank">http://sbml.org/Software/libSBML/docs/python-api/libsbml-accessing.htm</a> > > l#accessing-python > > > > Be sure to set properly 'DYLD_LIBRARY_PATH' as well. > > > >> > >> Nico > >> > >> > >> On 11/22/2012 12:34 AM, Lofgren, Eric wrote: > >>> I'm attempting to install libsbml with the accompanying Python bindings, > and am having a fair amount of trouble - and to be honest, I'm not sure quite > where things are going wrong. I'm using the Enthought Python Distribution, > which does admittedly add a layer of complexity to things. > >>> > >>> Following some advice from their listserv, I set up the configuration of > libsbml 5.6.0 using the following: > >>> > >>> ./configure > >>> --prefix=/Library/Frameworks/EPD64.framework/Versions/Current > >>> --with-python=/Library/Frameworks/EPD64.framework/Versions/Current > >>> > >>> Followed with: > >>> make > >>> and > >>> make install > >>> > >>> After that, I attempted to use libsbml's python bindings with the usual > from libsbml import * and got the following error: > >>> > >>>>>> from libsbml import * > >>> Traceback (most recent call last): > >>> File "", line 1, in > >>> ImportError: No module named libsbml > >>> > >>> Either libsbml isn't being installed correctly, or Python/EPD is having issues > finding the module. I can post the output from make and make install, but I > figured before I hit folks with a wall of text, I'd see if anyone had a solution > they knew of. > >>> > >>> Thanks, > >>> > >>> Eric > >>> ____________________________________________________________ > >>> To manage your sbml-interoperability list subscription, visit > >>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > >>> > >>> For a web interface to the sbml-interoperability mailing list, visit > >>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > >>> > >>> For questions or feedback about the sbml-interoperability list, > >>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > >> ____________________________________________________________ > >> To manage your sbml-interoperability list subscription, visit > >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > >> > >> For a web interface to the sbml-interoperability mailing list, visit > >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > >> > >> For questions or feedback about the sbml-interoperability list, > >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > > > ____________________________________________________________ > > To manage your sbml-interoperability list subscription, visit > > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > > > For a web interface to the sbml-interoperability mailing list, visit > > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > > > For questions or feedback about the sbml-interoperability list, > > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, contact sbml- > <a href="mailto:team@caltech.edu" target="_blank">team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Issues installing libsbml and Python bindings in OS X and EPD

rodrigue — 2012-11-23T10:24:54-00:00

On 11/22/2012 10:07 PM, Lofgren, Eric wrote: > This appears to have actually been an installation issue - searching for libsmbl.py only brings up the copies sitting in the downloaded libsbml-5.6.0 folder on my desktop, rather than where they should be in the OS X file system. That suggests to me that something's not installing correctly. > > I've run .configure along with make and make install - what terminal output would be useful in diagnosing an installation problem? did you use the --with-python option during configure ? see <a href="http://sbml.org/Software/libSBML/docs/python-api/libsbml-installation.html#unix-gnumake" target="_blank">http://sbml.org/Software/libSBML/docs/python-api/libsbml-installation.html#unix-gnumake</a> > > Eric > > On Nov 22, 2012, at 5:58 AM, Nicolas Rodriguez wrote: > >> On 11/22/2012 10:51 AM, Nicolas Rodriguez wrote: >>> This is probably a PYTHONPATH problem, search for the location of >>> libsbml.py and add the folder to the PYTHONPATH. >>> Then python will try to load the libsbml libraries and you might have >>> further problems. >> more infos can be found there : >> >> <a href="http://sbml.org/Software/libSBML/docs/python-api/libsbml-accessing.html#accessing-python" target="_blank">http://sbml.org/Software/libSBML/docs/python-api/libsbml-accessing.html#accessing-python</a> >> >> Be sure to set properly 'DYLD_LIBRARY_PATH' as well. >> >>> Nico >>> >>> >>> On 11/22/2012 12:34 AM, Lofgren, Eric wrote: >>>> I'm attempting to install libsbml with the accompanying Python bindings, and am having a fair amount of trouble - and to be honest, I'm not sure quite where things are going wrong. I'm using the Enthought Python Distribution, which does admittedly add a layer of complexity to things. >>>> >>>> Following some advice from their listserv, I set up the configuration of libsbml 5.6.0 using the following: >>>> >>>> ./configure --prefix=/Library/Frameworks/EPD64.framework/Versions/Current --with-python=/Library/Frameworks/EPD64.framework/Versions/Current >>>> >>>> Followed with: >>>> make >>>> and >>>> make install >>>> >>>> After that, I attempted to use libsbml's python bindings with the usual from libsbml import * and got the following error: >>>> >>>>>>> from libsbml import * >>>> Traceback (most recent call last): >>>> File "", line 1, in >>>> ImportError: No module named libsbml >>>> >>>> Either libsbml isn't being installed correctly, or Python/EPD is having issues finding the module. I can post the output from make and make install, but I figured before I hit folks with a wall of text, I'd see if anyone had a solution they knew of. >>>> >>>> Thanks, >>>> >>>> Eric >>>> ____________________________________________________________ >>>> To manage your sbml-interoperability list subscription, visit >>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >>>> >>>> For a web interface to the sbml-interoperability mailing list, visit >>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>> >>>> For questions or feedback about the sbml-interoperability list, >>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>> ____________________________________________________________ >>> To manage your sbml-interoperability list subscription, visit >>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >>> >>> For a web interface to the sbml-interoperability mailing list, visit >>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>> >>> For questions or feedback about the sbml-interoperability list, >>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >> ____________________________________________________________ >> To manage your sbml-interoperability list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >> >> For a web interface to the sbml-interoperability mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-interoperability list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Issues installing libsbml and Python bindings in OS X and EPD

elofgren — 2012-11-22T22:07:36-00:00

This appears to have actually been an installation issue - searching for libsmbl.py only brings up the copies sitting in the downloaded libsbml-5.6.0 folder on my desktop, rather than where they should be in the OS X file system. That suggests to me that something's not installing correctly. I've run .configure along with make and make install - what terminal output would be useful in diagnosing an installation problem? Eric On Nov 22, 2012, at 5:58 AM, Nicolas Rodriguez wrote: > On 11/22/2012 10:51 AM, Nicolas Rodriguez wrote: >> This is probably a PYTHONPATH problem, search for the location of >> libsbml.py and add the folder to the PYTHONPATH. >> Then python will try to load the libsbml libraries and you might have >> further problems. > > more infos can be found there : > > <a href="http://sbml.org/Software/libSBML/docs/python-api/libsbml-accessing.html#accessing-python" target="_blank">http://sbml.org/Software/libSBML/docs/python-api/libsbml-accessing.html#accessing-python</a> > > Be sure to set properly 'DYLD_LIBRARY_PATH' as well. > >> >> Nico >> >> >> On 11/22/2012 12:34 AM, Lofgren, Eric wrote: >>> I'm attempting to install libsbml with the accompanying Python bindings, and am having a fair amount of trouble - and to be honest, I'm not sure quite where things are going wrong. I'm using the Enthought Python Distribution, which does admittedly add a layer of complexity to things. >>> >>> Following some advice from their listserv, I set up the configuration of libsbml 5.6.0 using the following: >>> >>> ./configure --prefix=/Library/Frameworks/EPD64.framework/Versions/Current --with-python=/Library/Frameworks/EPD64.framework/Versions/Current >>> >>> Followed with: >>> make >>> and >>> make install >>> >>> After that, I attempted to use libsbml's python bindings with the usual from libsbml import * and got the following error: >>> >>>>>> from libsbml import * >>> Traceback (most recent call last): >>> File "", line 1, in >>> ImportError: No module named libsbml >>> >>> Either libsbml isn't being installed correctly, or Python/EPD is having issues finding the module. I can post the output from make and make install, but I figured before I hit folks with a wall of text, I'd see if anyone had a solution they knew of. >>> >>> Thanks, >>> >>> Eric >>> ____________________________________________________________ >>> To manage your sbml-interoperability list subscription, visit >>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >>> >>> For a web interface to the sbml-interoperability mailing list, visit >>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>> >>> For questions or feedback about the sbml-interoperability list, >>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >> ____________________________________________________________ >> To manage your sbml-interoperability list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >> >> For a web interface to the sbml-interoperability mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-interoperability list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Issues installing libsbml and Python bindings in OS X and EPD

rodrigue — 2012-11-22T10:58:57-00:00

On 11/22/2012 10:51 AM, Nicolas Rodriguez wrote: > This is probably a PYTHONPATH problem, search for the location of > libsbml.py and add the folder to the PYTHONPATH. > Then python will try to load the libsbml libraries and you might have > further problems. more infos can be found there : <a href="http://sbml.org/Software/libSBML/docs/python-api/libsbml-accessing.html#accessing-python" target="_blank">http://sbml.org/Software/libSBML/docs/python-api/libsbml-accessing.html#accessing-python</a> Be sure to set properly 'DYLD_LIBRARY_PATH' as well. > > Nico > > > On 11/22/2012 12:34 AM, Lofgren, Eric wrote: >> I'm attempting to install libsbml with the accompanying Python bindings, and am having a fair amount of trouble - and to be honest, I'm not sure quite where things are going wrong. I'm using the Enthought Python Distribution, which does admittedly add a layer of complexity to things. >> >> Following some advice from their listserv, I set up the configuration of libsbml 5.6.0 using the following: >> >> ./configure --prefix=/Library/Frameworks/EPD64.framework/Versions/Current --with-python=/Library/Frameworks/EPD64.framework/Versions/Current >> >> Followed with: >> make >> and >> make install >> >> After that, I attempted to use libsbml's python bindings with the usual from libsbml import * and got the following error: >> >>>>> from libsbml import * >> Traceback (most recent call last): >> File "", line 1, in >> ImportError: No module named libsbml >> >> Either libsbml isn't being installed correctly, or Python/EPD is having issues finding the module. I can post the output from make and make install, but I figured before I hit folks with a wall of text, I'd see if anyone had a solution they knew of. >> >> Thanks, >> >> Eric >> ____________________________________________________________ >> To manage your sbml-interoperability list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >> >> For a web interface to the sbml-interoperability mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-interoperability list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Issues installing libsbml and Python bindings in OS X and EPD

rodrigue — 2012-11-22T10:51:42-00:00

This is probably a PYTHONPATH problem, search for the location of libsbml.py and add the folder to the PYTHONPATH. Then python will try to load the libsbml libraries and you might have further problems. Nico On 11/22/2012 12:34 AM, Lofgren, Eric wrote: > I'm attempting to install libsbml with the accompanying Python bindings, and am having a fair amount of trouble - and to be honest, I'm not sure quite where things are going wrong. I'm using the Enthought Python Distribution, which does admittedly add a layer of complexity to things. > > Following some advice from their listserv, I set up the configuration of libsbml 5.6.0 using the following: > > ./configure --prefix=/Library/Frameworks/EPD64.framework/Versions/Current --with-python=/Library/Frameworks/EPD64.framework/Versions/Current > > Followed with: > make > and > make install > > After that, I attempted to use libsbml's python bindings with the usual from libsbml import * and got the following error: > >>>> from libsbml import * > Traceback (most recent call last): > File "", line 1, in > ImportError: No module named libsbml > > Either libsbml isn't being installed correctly, or Python/EPD is having issues finding the module. I can post the output from make and make install, but I figured before I hit folks with a wall of text, I'd see if anyone had a solution they knew of. > > Thanks, > > Eric > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Issues installing libsbml and Python bindings in OS X and EPD

elofgren — 2012-11-22T00:34:40-00:00

I'm attempting to install libsbml with the accompanying Python bindings, and am having a fair amount of trouble - and to be honest, I'm not sure quite where things are going wrong. I'm using the Enthought Python Distribution, which does admittedly add a layer of complexity to things. Following some advice from their listserv, I set up the configuration of libsbml 5.6.0 using the following: ./configure --prefix=/Library/Frameworks/EPD64.framework/Versions/Current --with-python=/Library/Frameworks/EPD64.framework/Versions/Current Followed with: make and make install After that, I attempted to use libsbml's python bindings with the usual from libsbml import * and got the following error: >>> from libsbml import * Traceback (most recent call last): File "", line 1, in ImportError: No module named libsbml Either libsbml isn't being installed correctly, or Python/EPD is having issues finding the module. I can post the output from make and make install, but I figured before I hit folks with a wall of text, I'd see if anyone had a solution they knew of. Thanks, Eric ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: retrieving model notes

mneal — 2012-11-15T12:27:49-00:00

Thanks, Frank. That helps very much. M --------------------------------- Maxwell Neal Post-doctoral researcher Department of Bioengineering University of Washington <a href="mailto:mneal@uw.edu" target="_blank">mneal@uw.edu</a> (206) 685-5173 --------------------------------- On Nov 15, 2012, at 4:22 AM, Frank T. Bergmann wrote: > Hello Max, > > sorry to hide this from you. The C API Documentation is still in the works. However, we made sure that most if not all functionality of libSBML is available for all languages. So until we have that documentation ready let me explain to you how you could reason out the functions you need: > > In C, we prefix the functions for an object with that objects name. So for example the counter part to your Java method Model.createEvent() would be Model_createEvent (Model_t *m). You don't have objects, so you have to pass the model object into each function. What about the inherited functions like the ones about notes that you ask for. Since you don't have objects in C, you won't have inheritance, so all attributes on SBML objects that are defined in the SBase class, will have to be prefixed by SBase_, so the function you originally asked for would have the signatures: > > SBase_getAnnotationString (SBase_t *sb); > > (Remember to cast your Model_t* to an SBase_t*!). > > I hope that helps, > > Cheers > Frank > > On Nov 14, 2012, at 11:22 PM, Maxwell Neal <<a href="mailto:mneal@u.washington.edu" target="_blank">mneal@u.washington.edu</a>> wrote: > >> Hi all, >> >> I'm working with libSBML in C and am looking for a way to retrieve the notes associated with a model. >> Previously, working with the java version (a while ago), I used sbmlmodel.getNotesString(). >> >> Is there a similar method available in C? >> Thanks, >> >> M >> >> --------------------------------- >> Maxwell Neal >> >> Post-doctoral researcher >> Department of Bioengineering >> University of Washington >> <a href="mailto:mneal@uw.edu" target="_blank">mneal@uw.edu</a> >> (206) 685-5173 >> --------------------------------- >> >> >> >> >> >> >> >> ____________________________________________________________ >> To manage your sbml-interoperability list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> >> >> For a web interface to the sbml-interoperability mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-interoperability list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: retrieving model notes

Sarah Keating — 2012-11-15T12:26:36-00:00

Hi Maxwell The getNotes functions in C are on the SBase structure and so will work with on all elements LIBSBML_EXTERN XMLNode_t * SBase_getNotes (SBase_t *sb); LIBSBML_EXTERN char* SBase_getNotesString (SBase_t *sb); Hope this helps :-) Sarah ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: retrieving model notes

fbergman — 2012-11-15T12:22:08-00:00

Hello Max, sorry to hide this from you. The C API Documentation is still in the works. However, we made sure that most if not all functionality of libSBML is available for all languages. So until we have that documentation ready let me explain to you how you could reason out the functions you need: In C, we prefix the functions for an object with that objects name. So for example the counter part to your Java method Model.createEvent() would be Model_createEvent (Model_t *m). You don't have objects, so you have to pass the model object into each function. What about the inherited functions like the ones about notes that you ask for. Since you don't have objects in C, you won't have inheritance, so all attributes on SBML objects that are defined in the SBase class, will have to be prefixed by SBase_, so the function you originally asked for would have the signatures: SBase_getAnnotationString (SBase_t *sb); (Remember to cast your Model_t* to an SBase_t*!). I hope that helps, Cheers Frank On Nov 14, 2012, at 11:22 PM, Maxwell Neal <<a href="mailto:mneal@u.washington.edu" target="_blank">mneal@u.washington.edu</a>> wrote: > Hi all, > > I'm working with libSBML in C and am looking for a way to retrieve the notes associated with a model. > Previously, working with the java version (a while ago), I used sbmlmodel.getNotesString(). > > Is there a similar method available in C? > Thanks, > > M > > --------------------------------- > Maxwell Neal > > Post-doctoral researcher > Department of Bioengineering > University of Washington > <a href="mailto:mneal@uw.edu" target="_blank">mneal@uw.edu</a> > (206) 685-5173 > --------------------------------- > > > > > > > > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

retrieving model notes

mneal — 2012-11-14T22:22:34-00:00

Hi all, I'm working with libSBML in C and am looking for a way to retrieve the notes associated with a model. Previously, working with the java version (a while ago), I used sbmlmodel.getNotesString(). Is there a similar method available in C? Thanks, M --------------------------------- Maxwell Neal Post-doctoral researcher Department of Bioengineering University of Washington <a href="mailto:mneal@uw.edu" target="_blank">mneal@uw.edu</a> (206) 685-5173 --------------------------------- ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: invalid character error

fbergman — 2012-10-31T20:54:57-00:00

Hello Max, While your question is specific to some calls of the libSBML API, the problem actually lies elsewhere. LibSBML just does what the SBML specification says, and tells you to pass in all documents encodes as UTF-8. So what seems to have happened in your case, is that whatever method pulled the UTF-8 from the BioModels Database to your local machine must have changed the encoding before passing it on to libSBML. This can happen very simply, for example if you used a java web service, or Java Strings to pull the model there might be a first point to look at, as Java internally uses UTF-16 (same goes for the .NET languages). Thats why there is special code in our language bindings to work on the conversion of the strings. Unfortunately your mail does not contain enough information as to where the error might lie. If the above does not apply, I would like to ask you to just write down the string that you would pass into readFromString and save it to file instead. That way you could more readily analyze what is going on. All the best Frank On Oct 31, 2012, at 5:22 PM, Maxwell Neal <<a href="mailto:mneal@u.washington.edu" target="_blank">mneal@u.washington.edu</a>> wrote: > Hi all, > > I'm trying to create a function in C that pulls down each curated model on BioModels as a string, loads it into an SBMLDocument_t object, and then performs some processing using libSBML. I've run into a problem that appears to be an encoding issue, but I'm not positive. After receiving the SBML text from the web services, I use "SBMLReader_readSBMLFromString" to create a new SBMLDocument. However, these documents always have an "...Invalid character in XML content" error. > > Based on the line numbers indicated in the error message, each time the culprit appears to be the accented e in Nicolas Le Novere's name. Does this mean that I need to convert the string received from BioModels into a char-friendly format in order to avoid these errors? > > Many thanks! > > M > --------------------------------- > Maxwell Neal > > Post-doctoral researcher > Department of Bioengineering > University of Washington > <a href="mailto:mneal@uw.edu" target="_blank">mneal@uw.edu</a> > (206) 685-5173 > --------------------------------- > > > > > > > > ____________________________________________________________ > To manage your sbml-interoperability list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> > > For a web interface to the sbml-interoperability mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-interoperability list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

invalid character error

mneal — 2012-10-31T16:22:14-00:00

Hi all, I'm trying to create a function in C that pulls down each curated model on BioModels as a string, loads it into an SBMLDocument_t object, and then performs some processing using libSBML. I've run into a problem that appears to be an encoding issue, but I'm not positive. After receiving the SBML text from the web services, I use "SBMLReader_readSBMLFromString" to create a new SBMLDocument. However, these documents always have an "...Invalid character in XML content" error. Based on the line numbers indicated in the error message, each time the culprit appears to be the accented e in Nicolas Le Novere's name. Does this mean that I need to convert the string received from BioModels into a char-friendly format in order to avoid these errors? Many thanks! M --------------------------------- Maxwell Neal Post-doctoral researcher Department of Bioengineering University of Washington <a href="mailto:mneal@uw.edu" target="_blank">mneal@uw.edu</a> (206) 685-5173 --------------------------------- ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

HARMONY 2013 date survey

Mike Hucka — 2012-10-24T19:02:56-00:00

Hello COMBINE communities, The University of Connecticut Health Center (UCHC) has kindly agreed to host HARMONY 2013, and they would like everyone to help choose the best dates. Please fill out this short survey: <a href="http://www.surveymonkey.com/s/QNNV6NX" target="_blank">http://www.surveymonkey.com/s/QNNV6NX</a> The deadline for submitting replies is November 1, 2012. Thank you. MH -- Mike Hucka, Ph.D. -- <a href="mailto:mhucka@caltech.edu" target="_blank">mhucka@caltech.edu</a> -- <a href="http://www.cds.caltech.edu/~mhucka" target="_blank">http://www.cds.caltech.edu/~mhucka</a> Computing and Mathematical Sciences, California Institute of Technology, Pasadena, California, USA -- Twitter: @mhucka -- Skype: michaelhucka -- You can give me anonymous feedback via <a href="http://tinyurl.com/mhuckafeedback" target="_blank">http://tinyurl.com/mhuckafeedback</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Questions about Installing libsbml-5.6.0 matlab bindings in Fedora17

Sarah Keating — 2012-09-30T11:20:35-00:00

Hi Zheng Thanks for you interest in libSBML and I'm sorry you are having a problem. It looks like your build was successful and so the issue may be with the matlab path. I would suggest you open matlab and add the /usr/local/lib directory to the path. Then go to the libsbml/bindings/matlab directory and run installSBML I'm hoping this will work :-) If it does not please do not hesitate to contact me and we can try something else. Thanks Sarah ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: How do you find software?

pabloar — 2012-09-28T22:37:54-00:00

Hi Mike, That's a good question. As I was looking into learning more about SBML-compatible software tools a few months ago, I found the resources in the SBML Software Guide extremely valuable. I used all three: the matrix, the summary and the showcase. I've had (more than once) the idea of providing the matrix in a more user-friendly format than it is currently available (e.g. pdf?). I find the scroll bar annoying, and I would prefer to have the whole table in a document that I could navigate more easily, or print out. BTW, I thought I would mention that today, as I was composing this post, I had real trouble finding the SBML Software Guide page(s), and actually I had to go back to a link I had in a personal reference document (at one point I thought the pages had been taken away). It was only a few minutes later that I found the link is at the bottom of the Downloads page, which might not be the most intuitive for those who are not familiar with the site. I hope this helps... Best regards, Pablo.

How do you find software?

Mike Hucka — 2012-09-26T01:47:25-00:00

Here's a general question for you all: When you want to find software for some purpose in computational systems biology, what do you do? How do you find it? Do you Google and wade through the results, or is there a software directory you use, or some other resource? (This has relevance for the SBML Team, because we keep the SBML Software Guide, but obviously there's more software out there besides the SBML-aware kind, and anyway, we don't know how much people use the Guide.) Thanks, MH -- Mike Hucka, Ph.D. -- <a href="mailto:mhucka@caltech.edu" target="_blank">mhucka@caltech.edu</a> -- <a href="http://www.cds.caltech.edu/~mhucka" target="_blank">http://www.cds.caltech.edu/~mhucka</a> Computing and Mathematical Sciences, California Institute of Technology, Pasadena, California, USA -- Twitter: @mhucka -- Skype: michaelhucka -- You can give me anonymous feedback via <a href="http://tinyurl.com/mhuckafeedback" target="_blank">http://tinyurl.com/mhuckafeedback</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Questions about Installing libsbml-5.6.0 matlab bindings in Fedora17

rz1991 — 2012-09-24T19:38:58-00:00

Hi everyone, I'm new to SBML and try to install libsbml-5.6.0 in my own laptop that runs Fedora17. I follows the documentation in the website telling me how to install it, but some problems just cannot be fixed myself. Hoping someone could give me a hand. I unpacked the source tar.gz and run the following command in the shell: > ./configure --with-matlab=/home/rz/Software/matlab --with-expat --with-python --with-python-interpreter=/usr/bin/python > make > sudo make install > env | grep "LD_LIBRARY_PATH" LD_LIBRARY_PATH=/usr/local/lib Everything seems correct. I checked the /usr/local/lib directory to see whether TranslateSBML.mexa64 and TraslateSBML.m is there and I get the positive result. Then I open the matlab2010a and make its working directory to ./src/bindings/matlab. By typing the following command in matlab environment: >> buildSBML * Doing preliminary checks of runtime environment ... - This appears to be MATLAB and not Octave. - MATLAB reports the OS is 64-bit Linux. Constructing the libSBML MATLAB interface. * Trying to establish our location ... - We are in the libSBML subdirectory for Matlab. - We appear to be in the libSBML source tree. - We have write access here! That makes us happy. * Locating libSBML library and include files ... - Found /home/rz/Software/libsbml-5.6.0/src/.libs/libsbml.so - Root of includes is /home/rz/Software/libsbml-5.6.0/src * Creating mex files in /home/rz/Software/libsbml-5.6.0/src/bindings/matlab - Building TranslateSBML ... Warning: You are using gcc version "4.7.0". The version currently supported with MEX is "4.2.3". For a list of currently supported compilers see: <a href="http://www.mathworks.com/support/compilers/current_release/" target="_blank">http://www.mathworks.com/support/compilers/current_release/</a> - Building OutputSBML ... Warning: You are using gcc version "4.7.0". The version currently supported with MEX is "4.2.3". For a list of currently supported compilers see: <a href="http://www.mathworks.com/support/compilers/current_release/" target="_blank">http://www.mathworks.com/support/compilers/current_release/</a> OutputSBML.c: In function ¡®mexFunction¡¯: OutputSBML.c:193:14: warning: assignment discards ¡®const¡¯ qualifier from pointer target type [enabled by default] Successfully finished. If appropriate, please run "installSBML" next. >> install Installing the libSBML MATLAB interface. * Doing preliminary checks of runtime environment ... - This appears to be MATLAB and not Octave. - MATLAB reports the OS is 64-bit Linux. Checking for executables ... Executables found * Trying to establish our location ... - We are in the libSBML subdirectory for Matlab. - We appear to be in the libSBML source tree. - We have write access here! That makes us happy. Checking for libraries ... adding /home/rz/Software/libsbml-5.6.0/src/bindings/matlab to the path Testing the installation. checking for TranslateSBML Installation failed - MATLAB cannot find all the libraries Add the path to the libraries to your path If this fails please contact the <a href="mailto:libsbml-team@caltech.edu" target="_blank">libsbml-team@caltech.edu</a> I was stuck here and know nothing to do. There is a warning by matlab that gcc4.7.0 is not well supported, but this is the default compiler of Fedora17. Is this the reason that I failed to install libsbml matlab binding? I hope someone could help me out. Thank you so much!! Zheng

Re: [sbml-distrib] Custom Annotation for drawing numbers from Distributions

Herbert M Sauro — 2012-09-07T15:12:53-00:00

Chris I think it is a good idea, in fact this what I going to put together to support missing functions I use, eg mod, may be a simple uniform random number generator like rand48() and some basic signals like sinusoidal, square wave, saw tooth, triangle and maybe others. Might be worth hosting such a library on sbml.org so every one has access. This might be much simpler than what you have in mind, the thing is I need some of these right now. Herbert Sent from my iPad On Sep 7, 2012, at 7:37 AM, "Chris J. Myers" <<a href="mailto:myers@ece.utah.edu" target="_blank">myers@ece.utah.edu</a>> wrote: > > By the way, what do people think of generalizing the package? I think we could add a definition URI to function definitions as a general scheme for creating new useful functions. This would mean this URI would be part of some "special function" package. The distrib package would then become simply a description of how the definition URI is used to create distributions and what subset of uncertML that we plan to support. Since the definition URI idea is generally useful, it seems odd to me to have multiple namespaces for different packages to add definition URIs when one namespace will do the trick. > > Cheers, > > Chris > > ---------------------------------- > Distributions we support: > > createFunction(model, "uniform", "Uniform distribution", "lambda(a,b,(a+b)/2)"); > createFunction(model, "normal", "Normal distribution", "lambda(m,s,m)"); > createFunction(model, "exponential", "Exponential distribution", "lambda(l,1/l)"); > createFunction(model, "gamma", "Gamma distribution", "lambda(a,b,a*b)"); > createFunction(model, "lognormal", "Lognormal distribution", "lambda(z,s,exp(z+s^2/2))"); > createFunction(model, "chisq", "Chi-squared distribution", "lambda(nu,nu)"); > createFunction(model, "laplace", "Laplace distribution", "lambda(a,0)"); > createFunction(model, "cauchy", "Cauchy distribution", "lambda(a,a)"); > createFunction(model, "rayleigh", "Rayleigh distribution","lambda(s,s*sqrt(pi/2))"); > createFunction(model, "poisson", "Poisson distribution", "lambda(mu,mu)"); > createFunction(model, "binomial", "Binomial distribution", "lambda(p,n,p*n)"); > createFunction(model, "bernoulli", "Bernoulli distribution", "lambda(p,p)"); > > > > On Sep 7, 2012, at 3:33 AM, Stuart Moodie wrote: > >> Hi Frank, >> >> On 7 Sep 2012, at 09:43, Frank T. Bergmann wrote: >> >>> In that case I shall await what happens to distrib with the new draft and continue to use the custom annotation for now. I still feel that distrib (and uncertML) do way more than what I wanted to accomplish for now (and what is implemented in available software). So it would feel wrong to use it as annotation when I would use a minuscule subset. >>> >> >> We had a chat with Dan Cornford from UncertML last week as I wanted to make sure that we could use UncertML as our definition of distributions. As Nicolas has said the idea - as in the current draft proposal - is to use a URI to define the distribution so your annotation is completely consistent with that. The outcome of the conversation is that UncertML will do what we want and that our working assumption is that the proposal will recommend the use of UncertML to define distributions. The benefits of UncertML are: >> >> 1) It is comprehensive. >> >> 2) It was designed for this purpose. >> >> 3) They are happy to work with the SBML community to ensure inter-operability with between both standards. >> >> 4) Each definition has *one* clearly stated mathematical definition of a PDF. >> >> 5) The arguments are clearly defined and have relatively clear typing. BTW. This is an issue looking at your implementation - I had no idea whether the second parameter was the stddev or variance of the distribution. >> >> 6) There is a parallel XML Schema definition (not used by distrib), which *may* be useful for implementations, but which also helps disambiguate the textual definition. >> >> There were a number of issues we will need to change in the latest proposal in order to use UncertML, but I think that might apply to any such resource: >> >> 1) UncertML takes vector arguments for mean and variance in some distributions, so we will need to be explicit in defining whether we want to use the vector or scalar form of an argument. >> >> 2) UncertML does not support truncation. We believe we can solve this by defining upper and lower bounds in distrib and it is then down to the sampling implementation to ensure that the truncation if correctly accounted for, >> >> There are a couple of issues which we would like UncertML to address: >> >> 1) We want to ensure that our URI points to a persistent identifier. In particular we would like a versioning strategy so that we can always be sure to get the exact definition we originally pointed to. We are discussing this with Dan. >> >> 2) This occurred to me later, but we need an explicit identifier for an argument if we are to define it's type (see above) therefore we need to come up with a convention that will be adhered by and maintained by both parties. >> >> I'm hoping to have a revised proposal incorporating these changes soon. >> >>> But I'll be happy to support what i can once it is ready. >> >> I'd encourage you to use UncertML as your experiences will help us with distrib. There is nothing like an implementation to identify problems and potentially better ways of doing things - as I'm sure you'll agree. >> >> Best wishes, >> >> Stuart. >> ____________________________________________________________ >> To manage your sbml-discuss list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >> >> For a web interface to the sbml-discuss mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-discuss list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: [sbml-discuss] [sbml-distrib] Custom Annotation for drawing numbers from Distributions

Herbert M Sauro — 2012-09-07T15:12:53-00:00

Chris I think it is a good idea, in fact this what I going to put together to support missing functions I use, eg mod, may be a simple uniform random number generator like rand48() and some basic signals like sinusoidal, square wave, saw tooth, triangle and maybe others. Might be worth hosting such a library on sbml.org so every one has access. This might be much simpler than what you have in mind, the thing is I need some of these right now. Herbert Sent from my iPad On Sep 7, 2012, at 7:37 AM, "Chris J. Myers" <<a href="mailto:myers@ece.utah.edu" target="_blank">myers@ece.utah.edu</a>> wrote: > > By the way, what do people think of generalizing the package? I think we could add a definition URI to function definitions as a general scheme for creating new useful functions. This would mean this URI would be part of some "special function" package. The distrib package would then become simply a description of how the definition URI is used to create distributions and what subset of uncertML that we plan to support. Since the definition URI idea is generally useful, it seems odd to me to have multiple namespaces for different packages to add definition URIs when one namespace will do the trick. > > Cheers, > > Chris > > ---------------------------------- > Distributions we support: > > createFunction(model, "uniform", "Uniform distribution", "lambda(a,b,(a+b)/2)"); > createFunction(model, "normal", "Normal distribution", "lambda(m,s,m)"); > createFunction(model, "exponential", "Exponential distribution", "lambda(l,1/l)"); > createFunction(model, "gamma", "Gamma distribution", "lambda(a,b,a*b)"); > createFunction(model, "lognormal", "Lognormal distribution", "lambda(z,s,exp(z+s^2/2))"); > createFunction(model, "chisq", "Chi-squared distribution", "lambda(nu,nu)"); > createFunction(model, "laplace", "Laplace distribution", "lambda(a,0)"); > createFunction(model, "cauchy", "Cauchy distribution", "lambda(a,a)"); > createFunction(model, "rayleigh", "Rayleigh distribution","lambda(s,s*sqrt(pi/2))"); > createFunction(model, "poisson", "Poisson distribution", "lambda(mu,mu)"); > createFunction(model, "binomial", "Binomial distribution", "lambda(p,n,p*n)"); > createFunction(model, "bernoulli", "Bernoulli distribution", "lambda(p,p)"); > > > > On Sep 7, 2012, at 3:33 AM, Stuart Moodie wrote: > >> Hi Frank, >> >> On 7 Sep 2012, at 09:43, Frank T. Bergmann wrote: >> >>> In that case I shall await what happens to distrib with the new draft and continue to use the custom annotation for now. I still feel that distrib (and uncertML) do way more than what I wanted to accomplish for now (and what is implemented in available software). So it would feel wrong to use it as annotation when I would use a minuscule subset. >>> >> >> We had a chat with Dan Cornford from UncertML last week as I wanted to make sure that we could use UncertML as our definition of distributions. As Nicolas has said the idea - as in the current draft proposal - is to use a URI to define the distribution so your annotation is completely consistent with that. The outcome of the conversation is that UncertML will do what we want and that our working assumption is that the proposal will recommend the use of UncertML to define distributions. The benefits of UncertML are: >> >> 1) It is comprehensive. >> >> 2) It was designed for this purpose. >> >> 3) They are happy to work with the SBML community to ensure inter-operability with between both standards. >> >> 4) Each definition has *one* clearly stated mathematical definition of a PDF. >> >> 5) The arguments are clearly defined and have relatively clear typing. BTW. This is an issue looking at your implementation - I had no idea whether the second parameter was the stddev or variance of the distribution. >> >> 6) There is a parallel XML Schema definition (not used by distrib), which *may* be useful for implementations, but which also helps disambiguate the textual definition. >> >> There were a number of issues we will need to change in the latest proposal in order to use UncertML, but I think that might apply to any such resource: >> >> 1) UncertML takes vector arguments for mean and variance in some distributions, so we will need to be explicit in defining whether we want to use the vector or scalar form of an argument. >> >> 2) UncertML does not support truncation. We believe we can solve this by defining upper and lower bounds in distrib and it is then down to the sampling implementation to ensure that the truncation if correctly accounted for, >> >> There are a couple of issues which we would like UncertML to address: >> >> 1) We want to ensure that our URI points to a persistent identifier. In particular we would like a versioning strategy so that we can always be sure to get the exact definition we originally pointed to. We are discussing this with Dan. >> >> 2) This occurred to me later, but we need an explicit identifier for an argument if we are to define it's type (see above) therefore we need to come up with a convention that will be adhered by and maintained by both parties. >> >> I'm hoping to have a revised proposal incorporating these changes soon. >> >>> But I'll be happy to support what i can once it is ready. >> >> I'd encourage you to use UncertML as your experiences will help us with distrib. There is nothing like an implementation to identify problems and potentially better ways of doing things - as I'm sure you'll agree. >> >> Best wishes, >> >> Stuart. >> ____________________________________________________________ >> To manage your sbml-discuss list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >> >> For a web interface to the sbml-discuss mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-discuss list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: [sbml-discuss] [sbml-distrib] Custom Annotation for drawing numbers from Distributions

myers — 2012-09-07T14:37:00-00:00

You mention changes need to the specification due to uncertML. Are you changing anything in the XML you sent earlier? Namely, do you need something more/different than a URI pointing to a definition? If not, shall we agree to use Frank's form with uncertML URIs for L2 and L3 before distrib is ready? Frank mentioned the distribution supported by iBioSim. We have mostly used the uniform and exponential distributions. However, we went ahead and added support for several more since it was easy to do given that we use the GNU Scientific Library. You can find the current list of distributions that GSL supports here: <a href="http://www.gnu.org/software/gsl/manual/html_node/Random-Number-Distributions.html" target="_blank">http://www.gnu.org/software/gsl/manual/html_node/Random-Number-Distributions.html</a> I think this list has grown recently because we actually only implemented a subset of this list (see below). An interesting point that Frank has raised I think is what does it mean to support the distrib package. Does a software have to support all possible distributions that we allow to be said to support the distrib package? If this is the case, we should likely be very selective about which distributions we include. On the other hand, we may then find people not being satisfied with the list and we are changing it all the time. On the other hand, we can try to provide a fairly exhaustive list while realizing that no software will likely support all the distributions on the list. This violates our notion of what it means for a package to be supported by software implementations. My view on this is that we should definitely not accept all of uncertML but do a subset like we do with mathML. Then, we should add terms whenever we can confirm that at least two tools will support a distribution. Since the goal is exchange, there must be the possibility of exchange. Note that this does not preclude though people from using the definition URI in a function for some function they want. However, it does limit what needs to be supported to state that you support this package. By the way, what do people think of generalizing the package? I think we could add a definition URI to function definitions as a general scheme for creating new useful functions. This would mean this URI would be part of some "special function" package. The distrib package would then become simply a description of how the definition URI is used to create distributions and what subset of uncertML that we plan to support. Since the definition URI idea is generally useful, it seems odd to me to have multiple namespaces for different packages to add definition URIs when one namespace will do the trick. Cheers, Chris ---------------------------------- Distributions we support: createFunction(model, "uniform", "Uniform distribution", "lambda(a,b,(a+b)/2)"); createFunction(model, "normal", "Normal distribution", "lambda(m,s,m)"); createFunction(model, "exponential", "Exponential distribution", "lambda(l,1/l)"); createFunction(model, "gamma", "Gamma distribution", "lambda(a,b,a*b)"); createFunction(model, "lognormal", "Lognormal distribution", "lambda(z,s,exp(z+s^2/2))"); createFunction(model, "chisq", "Chi-squared distribution", "lambda(nu,nu)"); createFunction(model, "laplace", "Laplace distribution", "lambda(a,0)"); createFunction(model, "cauchy", "Cauchy distribution", "lambda(a,a)"); createFunction(model, "rayleigh", "Rayleigh distribution","lambda(s,s*sqrt(pi/2))"); createFunction(model, "poisson", "Poisson distribution", "lambda(mu,mu)"); createFunction(model, "binomial", "Binomial distribution", "lambda(p,n,p*n)"); createFunction(model, "bernoulli", "Bernoulli distribution", "lambda(p,p)"); On Sep 7, 2012, at 3:33 AM, Stuart Moodie wrote: > Hi Frank, > > On 7 Sep 2012, at 09:43, Frank T. Bergmann wrote: > >> In that case I shall await what happens to distrib with the new draft and continue to use the custom annotation for now. I still feel that distrib (and uncertML) do way more than what I wanted to accomplish for now (and what is implemented in available software). So it would feel wrong to use it as annotation when I would use a minuscule subset. >> > > We had a chat with Dan Cornford from UncertML last week as I wanted to make sure that we could use UncertML as our definition of distributions. As Nicolas has said the idea - as in the current draft proposal - is to use a URI to define the distribution so your annotation is completely consistent with that. The outcome of the conversation is that UncertML will do what we want and that our working assumption is that the proposal will recommend the use of UncertML to define distributions. The benefits of UncertML are: > > 1) It is comprehensive. > > 2) It was designed for this purpose. > > 3) They are happy to work with the SBML community to ensure inter-operability with between both standards. > > 4) Each definition has *one* clearly stated mathematical definition of a PDF. > > 5) The arguments are clearly defined and have relatively clear typing. BTW. This is an issue looking at your implementation - I had no idea whether the second parameter was the stddev or variance of the distribution. > > 6) There is a parallel XML Schema definition (not used by distrib), which *may* be useful for implementations, but which also helps disambiguate the textual definition. > > There were a number of issues we will need to change in the latest proposal in order to use UncertML, but I think that might apply to any such resource: > > 1) UncertML takes vector arguments for mean and variance in some distributions, so we will need to be explicit in defining whether we want to use the vector or scalar form of an argument. > > 2) UncertML does not support truncation. We believe we can solve this by defining upper and lower bounds in distrib and it is then down to the sampling implementation to ensure that the truncation if correctly accounted for, > > There are a couple of issues which we would like UncertML to address: > > 1) We want to ensure that our URI points to a persistent identifier. In particular we would like a versioning strategy so that we can always be sure to get the exact definition we originally pointed to. We are discussing this with Dan. > > 2) This occurred to me later, but we need an explicit identifier for an argument if we are to define it's type (see above) therefore we need to come up with a convention that will be adhered by and maintained by both parties. > > I'm hoping to have a revised proposal incorporating these changes soon. > >> But I'll be happy to support what i can once it is ready. > > I'd encourage you to use UncertML as your experiences will help us with distrib. There is nothing like an implementation to identify problems and potentially better ways of doing things - as I'm sure you'll agree. > > Best wishes, > > Stuart. > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: [sbml-distrib] Custom Annotation for drawing numbers from Distributions

myers — 2012-09-07T14:37:00-00:00

You mention changes need to the specification due to uncertML. Are you changing anything in the XML you sent earlier? Namely, do you need something more/different than a URI pointing to a definition? If not, shall we agree to use Frank's form with uncertML URIs for L2 and L3 before distrib is ready? Frank mentioned the distribution supported by iBioSim. We have mostly used the uniform and exponential distributions. However, we went ahead and added support for several more since it was easy to do given that we use the GNU Scientific Library. You can find the current list of distributions that GSL supports here: <a href="http://www.gnu.org/software/gsl/manual/html_node/Random-Number-Distributions.html" target="_blank">http://www.gnu.org/software/gsl/manual/html_node/Random-Number-Distributions.html</a> I think this list has grown recently because we actually only implemented a subset of this list (see below). An interesting point that Frank has raised I think is what does it mean to support the distrib package. Does a software have to support all possible distributions that we allow to be said to support the distrib package? If this is the case, we should likely be very selective about which distributions we include. On the other hand, we may then find people not being satisfied with the list and we are changing it all the time. On the other hand, we can try to provide a fairly exhaustive list while realizing that no software will likely support all the distributions on the list. This violates our notion of what it means for a package to be supported by software implementations. My view on this is that we should definitely not accept all of uncertML but do a subset like we do with mathML. Then, we should add terms whenever we can confirm that at least two tools will support a distribution. Since the goal is exchange, there must be the possibility of exchange. Note that this does not preclude though people from using the definition URI in a function for some function they want. However, it does limit what needs to be supported to state that you support this package. By the way, what do people think of generalizing the package? I think we could add a definition URI to function definitions as a general scheme for creating new useful functions. This would mean this URI would be part of some "special function" package. The distrib package would then become simply a description of how the definition URI is used to create distributions and what subset of uncertML that we plan to support. Since the definition URI idea is generally useful, it seems odd to me to have multiple namespaces for different packages to add definition URIs when one namespace will do the trick. Cheers, Chris ---------------------------------- Distributions we support: createFunction(model, "uniform", "Uniform distribution", "lambda(a,b,(a+b)/2)"); createFunction(model, "normal", "Normal distribution", "lambda(m,s,m)"); createFunction(model, "exponential", "Exponential distribution", "lambda(l,1/l)"); createFunction(model, "gamma", "Gamma distribution", "lambda(a,b,a*b)"); createFunction(model, "lognormal", "Lognormal distribution", "lambda(z,s,exp(z+s^2/2))"); createFunction(model, "chisq", "Chi-squared distribution", "lambda(nu,nu)"); createFunction(model, "laplace", "Laplace distribution", "lambda(a,0)"); createFunction(model, "cauchy", "Cauchy distribution", "lambda(a,a)"); createFunction(model, "rayleigh", "Rayleigh distribution","lambda(s,s*sqrt(pi/2))"); createFunction(model, "poisson", "Poisson distribution", "lambda(mu,mu)"); createFunction(model, "binomial", "Binomial distribution", "lambda(p,n,p*n)"); createFunction(model, "bernoulli", "Bernoulli distribution", "lambda(p,p)"); On Sep 7, 2012, at 3:33 AM, Stuart Moodie wrote: > Hi Frank, > > On 7 Sep 2012, at 09:43, Frank T. Bergmann wrote: > >> In that case I shall await what happens to distrib with the new draft and continue to use the custom annotation for now. I still feel that distrib (and uncertML) do way more than what I wanted to accomplish for now (and what is implemented in available software). So it would feel wrong to use it as annotation when I would use a minuscule subset. >> > > We had a chat with Dan Cornford from UncertML last week as I wanted to make sure that we could use UncertML as our definition of distributions. As Nicolas has said the idea - as in the current draft proposal - is to use a URI to define the distribution so your annotation is completely consistent with that. The outcome of the conversation is that UncertML will do what we want and that our working assumption is that the proposal will recommend the use of UncertML to define distributions. The benefits of UncertML are: > > 1) It is comprehensive. > > 2) It was designed for this purpose. > > 3) They are happy to work with the SBML community to ensure inter-operability with between both standards. > > 4) Each definition has *one* clearly stated mathematical definition of a PDF. > > 5) The arguments are clearly defined and have relatively clear typing. BTW. This is an issue looking at your implementation - I had no idea whether the second parameter was the stddev or variance of the distribution. > > 6) There is a parallel XML Schema definition (not used by distrib), which *may* be useful for implementations, but which also helps disambiguate the textual definition. > > There were a number of issues we will need to change in the latest proposal in order to use UncertML, but I think that might apply to any such resource: > > 1) UncertML takes vector arguments for mean and variance in some distributions, so we will need to be explicit in defining whether we want to use the vector or scalar form of an argument. > > 2) UncertML does not support truncation. We believe we can solve this by defining upper and lower bounds in distrib and it is then down to the sampling implementation to ensure that the truncation if correctly accounted for, > > There are a couple of issues which we would like UncertML to address: > > 1) We want to ensure that our URI points to a persistent identifier. In particular we would like a versioning strategy so that we can always be sure to get the exact definition we originally pointed to. We are discussing this with Dan. > > 2) This occurred to me later, but we need an explicit identifier for an argument if we are to define it's type (see above) therefore we need to come up with a convention that will be adhered by and maintained by both parties. > > I'm hoping to have a revised proposal incorporating these changes soon. > >> But I'll be happy to support what i can once it is ready. > > I'd encourage you to use UncertML as your experiences will help us with distrib. There is nothing like an implementation to identify problems and potentially better ways of doing things - as I'm sure you'll agree. > > Best wishes, > > Stuart. > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

myers — 2012-09-07T14:11:52-00:00

I agree with Herbert that the seed is useful when you need reproducibility especially when you are debugging software. Chris On Sep 6, 2012, at 7:18 PM, hsauro wrote: > Assuming the algorithm is defined we should be able to store the seed > because sometimes we want to recreate a stochastic simulation from a > published paper in order to compute our own statistics about the simulation. > > Herbert > > On 9/6/2012 6:04 PM, Pedro Mendes wrote: >> >> On 09/07/2012 12:53 AM, Lucian Smith wrote: >>> In SBML, you only need a pointer that says 'this is a uniform random from >>> 0 (non-inclusive) to 1 (inclusive)' or the like. >> Exactly! The spec should define the behaviour of the function, not its >> implementation. No one in their right mind would like to define the >> cos() function as a series, but that is really how computers do it >> (sometimes combined with a lookup table to speed it up) >> >>> SED-ML would also be the place to store the random number seed, for >>> what small worth that gives you (i.e. only guaranteed on the exact >>> same machine that ran it the first time.) >> >> Seeding an RNG with a constant value is great for debugging code, >> otherwise it is definitely a bug since it will make your RND sequence >> completely predictable. The seed value should always be something >> unpredictable. So indicating the seed in SED-ML is really worth very >> little, I would even dare say: negative worth. >> >> -- >> Pedro Mendes >> Professor of Computational Systems Biology >> School of Computer Science >> Manchester Institute of Biotechnology >> The University of Manchester, >> 131 Princess St., Manchester, M1 7DN >> @gepasi <a href="http://www.comp-sys-bio.org" target="_blank">http://www.comp-sys-bio.org</a> >> ____________________________________________________________ >> To manage your sbml-discuss list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >> >> For a web interface to the sbml-discuss mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-discuss list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > > -- > Herbert M. Sauro, PhD > Associate Professor > Department of Bioengineering > > University of Washington > William H. Foege Building > Box 355061, Seattle, WA 98195-5061 > 206.685.2119 (tel), 206.685.3300 (shared fax) > > <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> > New Text Book: Enzyme Kinetics for Systems Biology > Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> > > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Model_add functions not updating model properly.

mihajlv — 2012-09-07T13:20:15-00:00

Thanks for letting me know, I'll give it a shot and let you know.

Re: [sbml-distrib] Custom Annotation for drawing numbers from Distributions

Stuart Moodie — 2012-09-07T09:33:13-00:00

Hi Frank, On 7 Sep 2012, at 09:43, Frank T. Bergmann wrote: > In that case I shall await what happens to distrib with the new draft and continue to use the custom annotation for now. I still feel that distrib (and uncertML) do way more than what I wanted to accomplish for now (and what is implemented in available software). So it would feel wrong to use it as annotation when I would use a minuscule subset. > We had a chat with Dan Cornford from UncertML last week as I wanted to make sure that we could use UncertML as our definition of distributions. As Nicolas has said the idea - as in the current draft proposal - is to use a URI to define the distribution so your annotation is completely consistent with that. The outcome of the conversation is that UncertML will do what we want and that our working assumption is that the proposal will recommend the use of UncertML to define distributions. The benefits of UncertML are: 1) It is comprehensive. 2) It was designed for this purpose. 3) They are happy to work with the SBML community to ensure inter-operability with between both standards. 4) Each definition has *one* clearly stated mathematical definition of a PDF. 5) The arguments are clearly defined and have relatively clear typing. BTW. This is an issue looking at your implementation - I had no idea whether the second parameter was the stddev or variance of the distribution. 6) There is a parallel XML Schema definition (not used by distrib), which *may* be useful for implementations, but which also helps disambiguate the textual definition. There were a number of issues we will need to change in the latest proposal in order to use UncertML, but I think that might apply to any such resource: 1) UncertML takes vector arguments for mean and variance in some distributions, so we will need to be explicit in defining whether we want to use the vector or scalar form of an argument. 2) UncertML does not support truncation. We believe we can solve this by defining upper and lower bounds in distrib and it is then down to the sampling implementation to ensure that the truncation if correctly accounted for, There are a couple of issues which we would like UncertML to address: 1) We want to ensure that our URI points to a persistent identifier. In particular we would like a versioning strategy so that we can always be sure to get the exact definition we originally pointed to. We are discussing this with Dan. 2) This occurred to me later, but we need an explicit identifier for an argument if we are to define it's type (see above) therefore we need to come up with a convention that will be adhered by and maintained by both parties. I'm hoping to have a revised proposal incorporating these changes soon. > But I'll be happy to support what i can once it is ready. I'd encourage you to use UncertML as your experiences will help us with distrib. There is nothing like an implementation to identify problems and potentially better ways of doing things - as I'm sure you'll agree. Best wishes, Stuart. ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: [sbml-distrib] [sbml-discuss] Custom Annotation for drawing numbers from Distributions

Stuart Moodie — 2012-09-07T09:33:13-00:00

Hi Frank, On 7 Sep 2012, at 09:43, Frank T. Bergmann wrote: > In that case I shall await what happens to distrib with the new draft and continue to use the custom annotation for now. I still feel that distrib (and uncertML) do way more than what I wanted to accomplish for now (and what is implemented in available software). So it would feel wrong to use it as annotation when I would use a minuscule subset. > We had a chat with Dan Cornford from UncertML last week as I wanted to make sure that we could use UncertML as our definition of distributions. As Nicolas has said the idea - as in the current draft proposal - is to use a URI to define the distribution so your annotation is completely consistent with that. The outcome of the conversation is that UncertML will do what we want and that our working assumption is that the proposal will recommend the use of UncertML to define distributions. The benefits of UncertML are: 1) It is comprehensive. 2) It was designed for this purpose. 3) They are happy to work with the SBML community to ensure inter-operability with between both standards. 4) Each definition has *one* clearly stated mathematical definition of a PDF. 5) The arguments are clearly defined and have relatively clear typing. BTW. This is an issue looking at your implementation - I had no idea whether the second parameter was the stddev or variance of the distribution. 6) There is a parallel XML Schema definition (not used by distrib), which *may* be useful for implementations, but which also helps disambiguate the textual definition. There were a number of issues we will need to change in the latest proposal in order to use UncertML, but I think that might apply to any such resource: 1) UncertML takes vector arguments for mean and variance in some distributions, so we will need to be explicit in defining whether we want to use the vector or scalar form of an argument. 2) UncertML does not support truncation. We believe we can solve this by defining upper and lower bounds in distrib and it is then down to the sampling implementation to ensure that the truncation if correctly accounted for, There are a couple of issues which we would like UncertML to address: 1) We want to ensure that our URI points to a persistent identifier. In particular we would like a versioning strategy so that we can always be sure to get the exact definition we originally pointed to. We are discussing this with Dan. 2) This occurred to me later, but we need an explicit identifier for an argument if we are to define it's type (see above) therefore we need to come up with a convention that will be adhered by and maintained by both parties. I'm hoping to have a revised proposal incorporating these changes soon. > But I'll be happy to support what i can once it is ready. I'd encourage you to use UncertML as your experiences will help us with distrib. There is nothing like an implementation to identify problems and potentially better ways of doing things - as I'm sure you'll agree. Best wishes, Stuart. ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: [sbml-discuss] Custom Annotation for drawing numbers from Distributions

fbergman — 2012-09-07T08:43:51-00:00

In that case I shall await what happens to distrib with the new draft and continue to use the custom annotation for now. I still feel that distrib (and uncertML) do way more than what I wanted to accomplish for now (and what is implemented in available software). So it would feel wrong to use it as annotation when I would use a minuscule subset. But I'll be happy to support what i can once it is ready. All the best Frank On Sep 7, 2012, at 10:31 AM, Nicolas Le Novère wrote: > On 07/09/12 09:22, Frank T. Bergmann wrote: >> That would solve one problem. However, a couple of practical issues remain >> that would make me want to hold on to the custom annotation (though I would >> be willing to also allow the other definition urls to be used): >> >> - distrib uses one attribute on the function definition to mark the >> function. this is difficult to put into a custom annotation (as i will need >> an element, so until and if distrib defines an L2 version I would be going >> on my own anyway) > > No this is gone. In the new iteration of the spec, the distrib constructs are on their own, which makes much easier to use them in L2. > >> - at this point i feel that uncertml goes beyond what I want to do, and >> really it would not help me. > > Our use of UncertML is 100% identical to your use of Wikipedia. It is a URI that points to a distribution definition. The difference for developers is that UncerML is defined in an XML schema, so you know the order of the arguments, and also you have an API which provides additional services. > >> Take the example that herbert posted yesterday: >> >> >> >> >> 3.14</un:logScale> >> 3.14</un:shape> >> </un:LogNormalDistribution> >> </distrib:distribution> >> >> >> mu</ci></bvar> >> sigma</ci></bvar> >> mu</ci> >> </lambda> >> </math> >> </functionDefinition> >> >> here the scale and shape are hardcoded with 3.14. Whereas what I wanted to >> do was to simply define a function definition to be used with the arguments >> I pass in later on. I'm sure this was discussed at COMBINE or HARMONY, and >> I just did not make it to that discussion. > > The hardcoded parameters are gone too. We now use UncertML as a controlled vocabulary, not an XML language anymore. > >> >> cheers >> Frank >> >> >> On Sep 7, 2012, at 10:08 AM, Nicolas Le Novère wrote: >> >>> Another draft is on the way. The URI will be or the form >>> <a href="http://www.uncertml.org/distributions/normal" target="_blank">http://www.uncertml.org/distributions/normal</a> >>> >>> We are working with UncertML to make sure that: >>> >>> * they cover everything we need >>> >>> If something is missing (e.g. something in iBioSim that is not covered), >>> we must tell them, directly or through the forum >>> <a href="http://uncertml.forum.52north.org/" target="_blank">http://uncertml.forum.52north.org/</a> >>> >>> * their URIs are stable >>> >>> Hopefully they are already. We are not their first user community >>> >>> * there is a versioning system in place >>> >>> On 07/09/12 08:59, Frank T. Bergmann wrote: >>>> Hello, >>>> >>>> what an amount of replies. I've picked this one to reply to because it >>>> most specifically applies to the original post. I index thought about >>>> the distrib proposal at first. However, it was much more detailed than >>>> what i wanted to cover at first. It also would not allow me to use the >>>> new functions in for example assignment rules. >>>> >>>> In the end it came down to the fact that i needed to support a certain >>>> number of functions, and there was no real identifier for them (the same >>>> is true of the distrib package, in section 3.7 of the draft where the >>>> distrib:function attribute value is described does not hold much >>>> information on which values would be allowed). The draft gets away with >>>> in the examples to use for example >>>> <a href="http://identifiers.org/truncatedNormal" target="_blank">http://identifiers.org/truncatedNormal</a> or >>>> "<a href="http:///identifiers.org/Distribution/LogNormal" target="_blank">http:///identifiers.org/Distribution/LogNormal</a>", but there are no such >>>> identifiers at the moment. At the same time the distributions listed >>>> under <a href="http://www.uncertml.org/distributions" target="_blank">http://www.uncertml.org/distributions</a> did not cover the same >>>> subset as for example supported by iBioSim. >>>> >>>> I don't mean to imply that the wikipedia entries are what is normative >>>> in this point or that they would have to be used programmatically. It is >>>> simply a helper to identify the kind of function i want, without having >>>> to hinge on the id / name of the function definition. For me the >>>> definition attribute in the custom annotation is too be seen just as >>>> well as the namespace identifier. >>>> >>>> Cheers >>>> Frank >>>> >>>> On Sep 6, 2012, at 6:15 PM, Nicolas Le Novère wrote: >>>> >>>>> Hi Frank, >>>>> >>>>> This is great that software start to exchange models with uncertainty. >>>>> But I am puzzled by your approach of developing your own annotation >>>>> rather than putting distrib constructs in the annotation element. That >>>>> would be the favoured way to provide an easier path towards the package >>>>> once accepted, and also minimise discrepancy between level 2 and level >>>>> 3. That is at least what has been done with the layout package. >>>>> >>>>> In effect what you propose is very close to distrib except you use >>>>> wikipedia instead of UncertML. However, since wikipedia cannot be used >>>>> easily programmaticaly, there is an uncertainty on the identity of the >>>>> parameter to feed the distribution (effectively the same problem we had >>>>> with the predefined functions of level 1). >>>>> >>>>> ? >>>>> >>>>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>>>> Hello everyone, >>>>>> >>>>>> I'm writing today to make you aware of a custom annotation for SBML >>>>>> Level 2 (or Level 3) models, that makes it possible to draw random >>>>>> numbers from a number of distributions. This is a feature many of us >>>>>> have encountered several times, but that is not yet standardized >>>>>> (until the release of the L3 distrib package). >>>>>> >>>>>> This is done by simply tagging a function definition with the right >>>>>> number of arguments, to let us know that this indeed is a distribution >>>>>> function. Here just one example: >>>>>> >>>>>> >>>>>> >>>>>> >>>>> definition="<a href="http://en.wikipedia.org/wiki/Normal_distribution" target="_blank">http://en.wikipedia.org/wiki/Normal_distribution</a>" /> >>>>>> </annotation> >>>>>> >>>>>> >>>>>> >>>>>> m </ci> >>>>>> </bvar> >>>>>> >>>>>> s </ci> >>>>>> </bvar> >>>>>> m </ci> >>>>>> </lambda> >>>>>> </math> >>>>>> </functionDefinition> >>>>>> >>>>>> By identifying the distribution function via the definition tag, it no >>>>>> longer matters how a tool names the function, making it easier to >>>>>> share across tools. Please let me know if you have any questions or >>>>>> concerns, or if you would like to implement support for it in your >>>>>> tool and require more information. >>>>>> >>>>>> more information and the specification here: >>>>>> - >>>>>> <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>>>> >>>>>> Cheers >>>>>> Frank >>>>>> ____________________________________________________________ >>>>>> To manage your sbml-discuss list subscription, visit >>>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>>> >>>>>> For a web interface to the sbml-discuss mailing list, visit >>>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>>> >>>>>> For questions or feedback about the sbml-discuss list, >>>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>>> >>>>> >>>>> >>>>> -- >>>>> Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, >>>>> Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, >>>>> <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere >>>>> <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> >>>>> >>>>> >>>> >>> >>> >>> -- >>> Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, >>> Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, >>> <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere >>> <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> >>> >>> >> > > > -- > Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, > Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, > <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere > <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> > > ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

fbergman — 2012-09-07T08:43:51-00:00

In that case I shall await what happens to distrib with the new draft and continue to use the custom annotation for now. I still feel that distrib (and uncertML) do way more than what I wanted to accomplish for now (and what is implemented in available software). So it would feel wrong to use it as annotation when I would use a minuscule subset. But I'll be happy to support what i can once it is ready. All the best Frank On Sep 7, 2012, at 10:31 AM, Nicolas Le Novère wrote: > On 07/09/12 09:22, Frank T. Bergmann wrote: >> That would solve one problem. However, a couple of practical issues remain >> that would make me want to hold on to the custom annotation (though I would >> be willing to also allow the other definition urls to be used): >> >> - distrib uses one attribute on the function definition to mark the >> function. this is difficult to put into a custom annotation (as i will need >> an element, so until and if distrib defines an L2 version I would be going >> on my own anyway) > > No this is gone. In the new iteration of the spec, the distrib constructs are on their own, which makes much easier to use them in L2. > >> - at this point i feel that uncertml goes beyond what I want to do, and >> really it would not help me. > > Our use of UncertML is 100% identical to your use of Wikipedia. It is a URI that points to a distribution definition. The difference for developers is that UncerML is defined in an XML schema, so you know the order of the arguments, and also you have an API which provides additional services. > >> Take the example that herbert posted yesterday: >> >> >> >> >> 3.14</un:logScale> >> 3.14</un:shape> >> </un:LogNormalDistribution> >> </distrib:distribution> >> >> >> mu</ci></bvar> >> sigma</ci></bvar> >> mu</ci> >> </lambda> >> </math> >> </functionDefinition> >> >> here the scale and shape are hardcoded with 3.14. Whereas what I wanted to >> do was to simply define a function definition to be used with the arguments >> I pass in later on. I'm sure this was discussed at COMBINE or HARMONY, and >> I just did not make it to that discussion. > > The hardcoded parameters are gone too. We now use UncertML as a controlled vocabulary, not an XML language anymore. > >> >> cheers >> Frank >> >> >> On Sep 7, 2012, at 10:08 AM, Nicolas Le Novère wrote: >> >>> Another draft is on the way. The URI will be or the form >>> <a href="http://www.uncertml.org/distributions/normal" target="_blank">http://www.uncertml.org/distributions/normal</a> >>> >>> We are working with UncertML to make sure that: >>> >>> * they cover everything we need >>> >>> If something is missing (e.g. something in iBioSim that is not covered), >>> we must tell them, directly or through the forum >>> <a href="http://uncertml.forum.52north.org/" target="_blank">http://uncertml.forum.52north.org/</a> >>> >>> * their URIs are stable >>> >>> Hopefully they are already. We are not their first user community >>> >>> * there is a versioning system in place >>> >>> On 07/09/12 08:59, Frank T. Bergmann wrote: >>>> Hello, >>>> >>>> what an amount of replies. I've picked this one to reply to because it >>>> most specifically applies to the original post. I index thought about >>>> the distrib proposal at first. However, it was much more detailed than >>>> what i wanted to cover at first. It also would not allow me to use the >>>> new functions in for example assignment rules. >>>> >>>> In the end it came down to the fact that i needed to support a certain >>>> number of functions, and there was no real identifier for them (the same >>>> is true of the distrib package, in section 3.7 of the draft where the >>>> distrib:function attribute value is described does not hold much >>>> information on which values would be allowed). The draft gets away with >>>> in the examples to use for example >>>> <a href="http://identifiers.org/truncatedNormal" target="_blank">http://identifiers.org/truncatedNormal</a> or >>>> "<a href="http:///identifiers.org/Distribution/LogNormal" target="_blank">http:///identifiers.org/Distribution/LogNormal</a>", but there are no such >>>> identifiers at the moment. At the same time the distributions listed >>>> under <a href="http://www.uncertml.org/distributions" target="_blank">http://www.uncertml.org/distributions</a> did not cover the same >>>> subset as for example supported by iBioSim. >>>> >>>> I don't mean to imply that the wikipedia entries are what is normative >>>> in this point or that they would have to be used programmatically. It is >>>> simply a helper to identify the kind of function i want, without having >>>> to hinge on the id / name of the function definition. For me the >>>> definition attribute in the custom annotation is too be seen just as >>>> well as the namespace identifier. >>>> >>>> Cheers >>>> Frank >>>> >>>> On Sep 6, 2012, at 6:15 PM, Nicolas Le Novère wrote: >>>> >>>>> Hi Frank, >>>>> >>>>> This is great that software start to exchange models with uncertainty. >>>>> But I am puzzled by your approach of developing your own annotation >>>>> rather than putting distrib constructs in the annotation element. That >>>>> would be the favoured way to provide an easier path towards the package >>>>> once accepted, and also minimise discrepancy between level 2 and level >>>>> 3. That is at least what has been done with the layout package. >>>>> >>>>> In effect what you propose is very close to distrib except you use >>>>> wikipedia instead of UncertML. However, since wikipedia cannot be used >>>>> easily programmaticaly, there is an uncertainty on the identity of the >>>>> parameter to feed the distribution (effectively the same problem we had >>>>> with the predefined functions of level 1). >>>>> >>>>> ? >>>>> >>>>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>>>> Hello everyone, >>>>>> >>>>>> I'm writing today to make you aware of a custom annotation for SBML >>>>>> Level 2 (or Level 3) models, that makes it possible to draw random >>>>>> numbers from a number of distributions. This is a feature many of us >>>>>> have encountered several times, but that is not yet standardized >>>>>> (until the release of the L3 distrib package). >>>>>> >>>>>> This is done by simply tagging a function definition with the right >>>>>> number of arguments, to let us know that this indeed is a distribution >>>>>> function. Here just one example: >>>>>> >>>>>> >>>>>> >>>>>> >>>>> definition="<a href="http://en.wikipedia.org/wiki/Normal_distribution" target="_blank">http://en.wikipedia.org/wiki/Normal_distribution</a>" /> >>>>>> </annotation> >>>>>> >>>>>> >>>>>> >>>>>> m </ci> >>>>>> </bvar> >>>>>> >>>>>> s </ci> >>>>>> </bvar> >>>>>> m </ci> >>>>>> </lambda> >>>>>> </math> >>>>>> </functionDefinition> >>>>>> >>>>>> By identifying the distribution function via the definition tag, it no >>>>>> longer matters how a tool names the function, making it easier to >>>>>> share across tools. Please let me know if you have any questions or >>>>>> concerns, or if you would like to implement support for it in your >>>>>> tool and require more information. >>>>>> >>>>>> more information and the specification here: >>>>>> - >>>>>> <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>>>> >>>>>> Cheers >>>>>> Frank >>>>>> ____________________________________________________________ >>>>>> To manage your sbml-discuss list subscription, visit >>>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>>> >>>>>> For a web interface to the sbml-discuss mailing list, visit >>>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>>> >>>>>> For questions or feedback about the sbml-discuss list, >>>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>>> >>>>> >>>>> >>>>> -- >>>>> Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, >>>>> Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, >>>>> <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere >>>>> <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> >>>>> >>>>> >>>> >>> >>> >>> -- >>> Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, >>> Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, >>> <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere >>> <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> >>> >>> >> > > > -- > Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, > Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, > <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere > <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> > > ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: [sbml-discuss] Custom Annotation for drawing numbers from Distributions

Nicolas Le Novere — 2012-09-07T08:31:11-00:00

On 07/09/12 09:22, Frank T. Bergmann wrote: > That would solve one problem. However, a couple of practical issues remain > that would make me want to hold on to the custom annotation (though I would > be willing to also allow the other definition urls to be used): > > - distrib uses one attribute on the function definition to mark the > function. this is difficult to put into a custom annotation (as i will need > an element, so until and if distrib defines an L2 version I would be going > on my own anyway) No this is gone. In the new iteration of the spec, the distrib constructs are on their own, which makes much easier to use them in L2. > - at this point i feel that uncertml goes beyond what I want to do, and > really it would not help me. Our use of UncertML is 100% identical to your use of Wikipedia. It is a URI that points to a distribution definition. The difference for developers is that UncerML is defined in an XML schema, so you know the order of the arguments, and also you have an API which provides additional services. > Take the example that herbert posted yesterday: > > > > > 3.14</un:logScale> > 3.14</un:shape> > </un:LogNormalDistribution> > </distrib:distribution> > > > mu</ci></bvar> > sigma</ci></bvar> > mu</ci> > </lambda> > </math> > </functionDefinition> > > here the scale and shape are hardcoded with 3.14. Whereas what I wanted to > do was to simply define a function definition to be used with the arguments > I pass in later on. I'm sure this was discussed at COMBINE or HARMONY, and > I just did not make it to that discussion. The hardcoded parameters are gone too. We now use UncertML as a controlled vocabulary, not an XML language anymore. > > cheers > Frank > > > On Sep 7, 2012, at 10:08 AM, Nicolas Le Novère wrote: > >> Another draft is on the way. The URI will be or the form >> <a href="http://www.uncertml.org/distributions/normal" target="_blank">http://www.uncertml.org/distributions/normal</a> >> >> We are working with UncertML to make sure that: >> >> * they cover everything we need >> >> If something is missing (e.g. something in iBioSim that is not covered), >> we must tell them, directly or through the forum >> <a href="http://uncertml.forum.52north.org/" target="_blank">http://uncertml.forum.52north.org/</a> >> >> * their URIs are stable >> >> Hopefully they are already. We are not their first user community >> >> * there is a versioning system in place >> >> On 07/09/12 08:59, Frank T. Bergmann wrote: >>> Hello, >>> >>> what an amount of replies. I've picked this one to reply to because it >>> most specifically applies to the original post. I index thought about >>> the distrib proposal at first. However, it was much more detailed than >>> what i wanted to cover at first. It also would not allow me to use the >>> new functions in for example assignment rules. >>> >>> In the end it came down to the fact that i needed to support a certain >>> number of functions, and there was no real identifier for them (the same >>> is true of the distrib package, in section 3.7 of the draft where the >>> distrib:function attribute value is described does not hold much >>> information on which values would be allowed). The draft gets away with >>> in the examples to use for example >>> <a href="http://identifiers.org/truncatedNormal" target="_blank">http://identifiers.org/truncatedNormal</a> or >>> "<a href="http:///identifiers.org/Distribution/LogNormal" target="_blank">http:///identifiers.org/Distribution/LogNormal</a>", but there are no such >>> identifiers at the moment. At the same time the distributions listed >>> under <a href="http://www.uncertml.org/distributions" target="_blank">http://www.uncertml.org/distributions</a> did not cover the same >>> subset as for example supported by iBioSim. >>> >>> I don't mean to imply that the wikipedia entries are what is normative >>> in this point or that they would have to be used programmatically. It is >>> simply a helper to identify the kind of function i want, without having >>> to hinge on the id / name of the function definition. For me the >>> definition attribute in the custom annotation is too be seen just as >>> well as the namespace identifier. >>> >>> Cheers >>> Frank >>> >>> On Sep 6, 2012, at 6:15 PM, Nicolas Le Novère wrote: >>> >>>> Hi Frank, >>>> >>>> This is great that software start to exchange models with uncertainty. >>>> But I am puzzled by your approach of developing your own annotation >>>> rather than putting distrib constructs in the annotation element. That >>>> would be the favoured way to provide an easier path towards the package >>>> once accepted, and also minimise discrepancy between level 2 and level >>>> 3. That is at least what has been done with the layout package. >>>> >>>> In effect what you propose is very close to distrib except you use >>>> wikipedia instead of UncertML. However, since wikipedia cannot be used >>>> easily programmaticaly, there is an uncertainty on the identity of the >>>> parameter to feed the distribution (effectively the same problem we had >>>> with the predefined functions of level 1). >>>> >>>> ? >>>> >>>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>>> Hello everyone, >>>>> >>>>> I'm writing today to make you aware of a custom annotation for SBML >>>>> Level 2 (or Level 3) models, that makes it possible to draw random >>>>> numbers from a number of distributions. This is a feature many of us >>>>> have encountered several times, but that is not yet standardized >>>>> (until the release of the L3 distrib package). >>>>> >>>>> This is done by simply tagging a function definition with the right >>>>> number of arguments, to let us know that this indeed is a distribution >>>>> function. Here just one example: >>>>> >>>>> >>>>> >>>>> >>>> definition="<a href="http://en.wikipedia.org/wiki/Normal_distribution" target="_blank">http://en.wikipedia.org/wiki/Normal_distribution</a>" /> >>>>> </annotation> >>>>> >>>>> >>>>> >>>>> m </ci> >>>>> </bvar> >>>>> >>>>> s </ci> >>>>> </bvar> >>>>> m </ci> >>>>> </lambda> >>>>> </math> >>>>> </functionDefinition> >>>>> >>>>> By identifying the distribution function via the definition tag, it no >>>>> longer matters how a tool names the function, making it easier to >>>>> share across tools. Please let me know if you have any questions or >>>>> concerns, or if you would like to implement support for it in your >>>>> tool and require more information. >>>>> >>>>> more information and the specification here: >>>>> - >>>>> <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>>> >>>>> Cheers >>>>> Frank >>>>> ____________________________________________________________ >>>>> To manage your sbml-discuss list subscription, visit >>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>> >>>>> For a web interface to the sbml-discuss mailing list, visit >>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>> >>>>> For questions or feedback about the sbml-discuss list, >>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>> >>>> >>>> >>>> -- >>>> Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, >>>> Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, >>>> <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere >>>> <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> >>>> >>>> >>> >> >> >> -- >> Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, >> Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, >> <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere >> <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> >> >> > -- Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

Nicolas Le Novere — 2012-09-07T08:31:11-00:00

On 07/09/12 09:22, Frank T. Bergmann wrote: > That would solve one problem. However, a couple of practical issues remain > that would make me want to hold on to the custom annotation (though I would > be willing to also allow the other definition urls to be used): > > - distrib uses one attribute on the function definition to mark the > function. this is difficult to put into a custom annotation (as i will need > an element, so until and if distrib defines an L2 version I would be going > on my own anyway) No this is gone. In the new iteration of the spec, the distrib constructs are on their own, which makes much easier to use them in L2. > - at this point i feel that uncertml goes beyond what I want to do, and > really it would not help me. Our use of UncertML is 100% identical to your use of Wikipedia. It is a URI that points to a distribution definition. The difference for developers is that UncerML is defined in an XML schema, so you know the order of the arguments, and also you have an API which provides additional services. > Take the example that herbert posted yesterday: > > > > > 3.14</un:logScale> > 3.14</un:shape> > </un:LogNormalDistribution> > </distrib:distribution> > > > mu</ci></bvar> > sigma</ci></bvar> > mu</ci> > </lambda> > </math> > </functionDefinition> > > here the scale and shape are hardcoded with 3.14. Whereas what I wanted to > do was to simply define a function definition to be used with the arguments > I pass in later on. I'm sure this was discussed at COMBINE or HARMONY, and > I just did not make it to that discussion. The hardcoded parameters are gone too. We now use UncertML as a controlled vocabulary, not an XML language anymore. > > cheers > Frank > > > On Sep 7, 2012, at 10:08 AM, Nicolas Le Novère wrote: > >> Another draft is on the way. The URI will be or the form >> <a href="http://www.uncertml.org/distributions/normal" target="_blank">http://www.uncertml.org/distributions/normal</a> >> >> We are working with UncertML to make sure that: >> >> * they cover everything we need >> >> If something is missing (e.g. something in iBioSim that is not covered), >> we must tell them, directly or through the forum >> <a href="http://uncertml.forum.52north.org/" target="_blank">http://uncertml.forum.52north.org/</a> >> >> * their URIs are stable >> >> Hopefully they are already. We are not their first user community >> >> * there is a versioning system in place >> >> On 07/09/12 08:59, Frank T. Bergmann wrote: >>> Hello, >>> >>> what an amount of replies. I've picked this one to reply to because it >>> most specifically applies to the original post. I index thought about >>> the distrib proposal at first. However, it was much more detailed than >>> what i wanted to cover at first. It also would not allow me to use the >>> new functions in for example assignment rules. >>> >>> In the end it came down to the fact that i needed to support a certain >>> number of functions, and there was no real identifier for them (the same >>> is true of the distrib package, in section 3.7 of the draft where the >>> distrib:function attribute value is described does not hold much >>> information on which values would be allowed). The draft gets away with >>> in the examples to use for example >>> <a href="http://identifiers.org/truncatedNormal" target="_blank">http://identifiers.org/truncatedNormal</a> or >>> "<a href="http:///identifiers.org/Distribution/LogNormal" target="_blank">http:///identifiers.org/Distribution/LogNormal</a>", but there are no such >>> identifiers at the moment. At the same time the distributions listed >>> under <a href="http://www.uncertml.org/distributions" target="_blank">http://www.uncertml.org/distributions</a> did not cover the same >>> subset as for example supported by iBioSim. >>> >>> I don't mean to imply that the wikipedia entries are what is normative >>> in this point or that they would have to be used programmatically. It is >>> simply a helper to identify the kind of function i want, without having >>> to hinge on the id / name of the function definition. For me the >>> definition attribute in the custom annotation is too be seen just as >>> well as the namespace identifier. >>> >>> Cheers >>> Frank >>> >>> On Sep 6, 2012, at 6:15 PM, Nicolas Le Novère wrote: >>> >>>> Hi Frank, >>>> >>>> This is great that software start to exchange models with uncertainty. >>>> But I am puzzled by your approach of developing your own annotation >>>> rather than putting distrib constructs in the annotation element. That >>>> would be the favoured way to provide an easier path towards the package >>>> once accepted, and also minimise discrepancy between level 2 and level >>>> 3. That is at least what has been done with the layout package. >>>> >>>> In effect what you propose is very close to distrib except you use >>>> wikipedia instead of UncertML. However, since wikipedia cannot be used >>>> easily programmaticaly, there is an uncertainty on the identity of the >>>> parameter to feed the distribution (effectively the same problem we had >>>> with the predefined functions of level 1). >>>> >>>> ? >>>> >>>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>>> Hello everyone, >>>>> >>>>> I'm writing today to make you aware of a custom annotation for SBML >>>>> Level 2 (or Level 3) models, that makes it possible to draw random >>>>> numbers from a number of distributions. This is a feature many of us >>>>> have encountered several times, but that is not yet standardized >>>>> (until the release of the L3 distrib package). >>>>> >>>>> This is done by simply tagging a function definition with the right >>>>> number of arguments, to let us know that this indeed is a distribution >>>>> function. Here just one example: >>>>> >>>>> >>>>> >>>>> >>>> definition="<a href="http://en.wikipedia.org/wiki/Normal_distribution" target="_blank">http://en.wikipedia.org/wiki/Normal_distribution</a>" /> >>>>> </annotation> >>>>> >>>>> >>>>> >>>>> m </ci> >>>>> </bvar> >>>>> >>>>> s </ci> >>>>> </bvar> >>>>> m </ci> >>>>> </lambda> >>>>> </math> >>>>> </functionDefinition> >>>>> >>>>> By identifying the distribution function via the definition tag, it no >>>>> longer matters how a tool names the function, making it easier to >>>>> share across tools. Please let me know if you have any questions or >>>>> concerns, or if you would like to implement support for it in your >>>>> tool and require more information. >>>>> >>>>> more information and the specification here: >>>>> - >>>>> <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>>> >>>>> Cheers >>>>> Frank >>>>> ____________________________________________________________ >>>>> To manage your sbml-discuss list subscription, visit >>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>> >>>>> For a web interface to the sbml-discuss mailing list, visit >>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>> >>>>> For questions or feedback about the sbml-discuss list, >>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>> >>>> >>>> >>>> -- >>>> Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, >>>> Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, >>>> <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere >>>> <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> >>>> >>>> >>> >> >> >> -- >> Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, >> Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, >> <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere >> <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> >> >> > -- Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: [sbml-discuss] Custom Annotation for drawing numbers from Distributions

fbergman — 2012-09-07T08:22:09-00:00

That would solve one problem. However, a couple of practical issues remain that would make me want to hold on to the custom annotation (though I would be willing to also allow the other definition urls to be used): - distrib uses one attribute on the function definition to mark the function. this is difficult to put into a custom annotation (as i will need an element, so until and if distrib defines an L2 version I would be going on my own anyway) - at this point i feel that uncertml goes beyond what I want to do, and really it would not help me. Take the example that herbert posted yesterday: 3.14</un:logScale> 3.14</un:shape> </un:LogNormalDistribution> </distrib:distribution> mu</ci></bvar> sigma</ci></bvar> mu</ci> </lambda> </math> </functionDefinition> here the scale and shape are hardcoded with 3.14. Whereas what I wanted to do was to simply define a function definition to be used with the arguments I pass in later on. I'm sure this was discussed at COMBINE or HARMONY, and I just did not make it to that discussion. cheers Frank On Sep 7, 2012, at 10:08 AM, Nicolas Le Novère wrote: > Another draft is on the way. The URI will be or the form > <a href="http://www.uncertml.org/distributions/normal" target="_blank">http://www.uncertml.org/distributions/normal</a> > > We are working with UncertML to make sure that: > > * they cover everything we need > > If something is missing (e.g. something in iBioSim that is not covered), we must tell them, directly or through the forum <a href="http://uncertml.forum.52north.org/" target="_blank">http://uncertml.forum.52north.org/</a> > > * their URIs are stable > > Hopefully they are already. We are not their first user community > > * there is a versioning system in place > > On 07/09/12 08:59, Frank T. Bergmann wrote: >> Hello, >> >> what an amount of replies. I've picked this one to reply to because it most specifically applies to the original post. I index thought about the distrib proposal at first. However, it was much more detailed than what i wanted to cover at first. It also would not allow me to use the new functions in for example assignment rules. >> >> In the end it came down to the fact that i needed to support a certain number of functions, and there was no real identifier for them (the same is true of the distrib package, in section 3.7 of the draft where the distrib:function attribute value is described does not hold much information on which values would be allowed). The draft gets away with in the examples to use for example <a href="http://identifiers.org/truncatedNormal" target="_blank">http://identifiers.org/truncatedNormal</a> or "<a href="http:///identifiers.org/Distribution/LogNormal" target="_blank">http:///identifiers.org/Distribution/LogNormal</a>", but there are no such identifiers at the moment. At the same time the distributions listed under <a href="http://www.uncertml.org/distributions" target="_blank">http://www.uncertml.org/distributions</a> did not cover the same subset as for example supported by iBioSim. >> >> I don't mean to imply that the wikipedia entries are what is normative in this point or that they would have to be used programmatically. It is simply a helper to identify the kind of function i want, without having to hinge on the id / name of the function definition. For me the definition attribute in the custom annotation is too be seen just as well as the namespace identifier. >> >> Cheers >> Frank >> >> On Sep 6, 2012, at 6:15 PM, Nicolas Le Novère wrote: >> >>> Hi Frank, >>> >>> This is great that software start to exchange models with uncertainty. But I am puzzled by your approach of developing your own annotation rather than putting distrib constructs in the annotation element. That would be the favoured way to provide an easier path towards the package once accepted, and also minimise discrepancy between level 2 and level 3. That is at least what has been done with the layout package. >>> >>> In effect what you propose is very close to distrib except you use wikipedia instead of UncertML. However, since wikipedia cannot be used easily programmaticaly, there is an uncertainty on the identity of the parameter to feed the distribution (effectively the same problem we had with the predefined functions of level 1). >>> >>> ? >>> >>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>> Hello everyone, >>>> >>>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >>>> >>>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >>>> >>>> >>>> >>>> >>>> </annotation> >>>> >>>> >>>> >>>> m </ci> >>>> </bvar> >>>> >>>> s </ci> >>>> </bvar> >>>> m </ci> >>>> </lambda> >>>> </math> >>>> </functionDefinition> >>>> >>>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >>>> >>>> more information and the specification here: >>>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>> >>>> Cheers >>>> Frank >>>> ____________________________________________________________ >>>> To manage your sbml-discuss list subscription, visit >>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>> >>>> For a web interface to the sbml-discuss mailing list, visit >>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>> >>>> For questions or feedback about the sbml-discuss list, >>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>> >>> >>> >>> -- >>> Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, >>> Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, >>> <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere >>> <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> >>> >>> >> > > > -- > Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, > Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, > <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere > <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> > > ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

fbergman — 2012-09-07T08:22:09-00:00

That would solve one problem. However, a couple of practical issues remain that would make me want to hold on to the custom annotation (though I would be willing to also allow the other definition urls to be used): - distrib uses one attribute on the function definition to mark the function. this is difficult to put into a custom annotation (as i will need an element, so until and if distrib defines an L2 version I would be going on my own anyway) - at this point i feel that uncertml goes beyond what I want to do, and really it would not help me. Take the example that herbert posted yesterday: 3.14</un:logScale> 3.14</un:shape> </un:LogNormalDistribution> </distrib:distribution> mu</ci></bvar> sigma</ci></bvar> mu</ci> </lambda> </math> </functionDefinition> here the scale and shape are hardcoded with 3.14. Whereas what I wanted to do was to simply define a function definition to be used with the arguments I pass in later on. I'm sure this was discussed at COMBINE or HARMONY, and I just did not make it to that discussion. cheers Frank On Sep 7, 2012, at 10:08 AM, Nicolas Le Novère wrote: > Another draft is on the way. The URI will be or the form > <a href="http://www.uncertml.org/distributions/normal" target="_blank">http://www.uncertml.org/distributions/normal</a> > > We are working with UncertML to make sure that: > > * they cover everything we need > > If something is missing (e.g. something in iBioSim that is not covered), we must tell them, directly or through the forum <a href="http://uncertml.forum.52north.org/" target="_blank">http://uncertml.forum.52north.org/</a> > > * their URIs are stable > > Hopefully they are already. We are not their first user community > > * there is a versioning system in place > > On 07/09/12 08:59, Frank T. Bergmann wrote: >> Hello, >> >> what an amount of replies. I've picked this one to reply to because it most specifically applies to the original post. I index thought about the distrib proposal at first. However, it was much more detailed than what i wanted to cover at first. It also would not allow me to use the new functions in for example assignment rules. >> >> In the end it came down to the fact that i needed to support a certain number of functions, and there was no real identifier for them (the same is true of the distrib package, in section 3.7 of the draft where the distrib:function attribute value is described does not hold much information on which values would be allowed). The draft gets away with in the examples to use for example <a href="http://identifiers.org/truncatedNormal" target="_blank">http://identifiers.org/truncatedNormal</a> or "<a href="http:///identifiers.org/Distribution/LogNormal" target="_blank">http:///identifiers.org/Distribution/LogNormal</a>", but there are no such identifiers at the moment. At the same time the distributions listed under <a href="http://www.uncertml.org/distributions" target="_blank">http://www.uncertml.org/distributions</a> did not cover the same subset as for example supported by iBioSim. >> >> I don't mean to imply that the wikipedia entries are what is normative in this point or that they would have to be used programmatically. It is simply a helper to identify the kind of function i want, without having to hinge on the id / name of the function definition. For me the definition attribute in the custom annotation is too be seen just as well as the namespace identifier. >> >> Cheers >> Frank >> >> On Sep 6, 2012, at 6:15 PM, Nicolas Le Novère wrote: >> >>> Hi Frank, >>> >>> This is great that software start to exchange models with uncertainty. But I am puzzled by your approach of developing your own annotation rather than putting distrib constructs in the annotation element. That would be the favoured way to provide an easier path towards the package once accepted, and also minimise discrepancy between level 2 and level 3. That is at least what has been done with the layout package. >>> >>> In effect what you propose is very close to distrib except you use wikipedia instead of UncertML. However, since wikipedia cannot be used easily programmaticaly, there is an uncertainty on the identity of the parameter to feed the distribution (effectively the same problem we had with the predefined functions of level 1). >>> >>> ? >>> >>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>> Hello everyone, >>>> >>>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >>>> >>>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >>>> >>>> >>>> >>>> >>>> </annotation> >>>> >>>> >>>> >>>> m </ci> >>>> </bvar> >>>> >>>> s </ci> >>>> </bvar> >>>> m </ci> >>>> </lambda> >>>> </math> >>>> </functionDefinition> >>>> >>>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >>>> >>>> more information and the specification here: >>>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>> >>>> Cheers >>>> Frank >>>> ____________________________________________________________ >>>> To manage your sbml-discuss list subscription, visit >>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>> >>>> For a web interface to the sbml-discuss mailing list, visit >>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>> >>>> For questions or feedback about the sbml-discuss list, >>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>> >>> >>> >>> -- >>> Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, >>> Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, >>> <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere >>> <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> >>> >>> >> > > > -- > Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, > Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, > <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere > <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> > > ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: [sbml-discuss] Custom Annotation for drawing numbers from Distributions

Nicolas Le Novere — 2012-09-07T08:08:11-00:00

Another draft is on the way. The URI will be or the form <a href="http://www.uncertml.org/distributions/normal" target="_blank">http://www.uncertml.org/distributions/normal</a> We are working with UncertML to make sure that: * they cover everything we need If something is missing (e.g. something in iBioSim that is not covered), we must tell them, directly or through the forum <a href="http://uncertml.forum.52north.org/" target="_blank">http://uncertml.forum.52north.org/</a> * their URIs are stable Hopefully they are already. We are not their first user community * there is a versioning system in place On 07/09/12 08:59, Frank T. Bergmann wrote: > Hello, > > what an amount of replies. I've picked this one to reply to because it most specifically applies to the original post. I index thought about the distrib proposal at first. However, it was much more detailed than what i wanted to cover at first. It also would not allow me to use the new functions in for example assignment rules. > > In the end it came down to the fact that i needed to support a certain number of functions, and there was no real identifier for them (the same is true of the distrib package, in section 3.7 of the draft where the distrib:function attribute value is described does not hold much information on which values would be allowed). The draft gets away with in the examples to use for example <a href="http://identifiers.org/truncatedNormal" target="_blank">http://identifiers.org/truncatedNormal</a> or "<a href="http:///identifiers.org/Distribution/LogNormal" target="_blank">http:///identifiers.org/Distribution/LogNormal</a>", but there are no such identifiers at the moment. At the same time the distributions listed under <a href="http://www.uncertml.org/distributions" target="_blank">http://www.uncertml.org/distributions</a> did not cover the same subset as for example supported by iBioSim. > > I don't mean to imply that the wikipedia entries are what is normative in this point or that they would have to be used programmatically. It is simply a helper to identify the kind of function i want, without having to hinge on the id / name of the function definition. For me the definition attribute in the custom annotation is too be seen just as well as the namespace identifier. > > Cheers > Frank > > On Sep 6, 2012, at 6:15 PM, Nicolas Le Novère wrote: > >> Hi Frank, >> >> This is great that software start to exchange models with uncertainty. But I am puzzled by your approach of developing your own annotation rather than putting distrib constructs in the annotation element. That would be the favoured way to provide an easier path towards the package once accepted, and also minimise discrepancy between level 2 and level 3. That is at least what has been done with the layout package. >> >> In effect what you propose is very close to distrib except you use wikipedia instead of UncertML. However, since wikipedia cannot be used easily programmaticaly, there is an uncertainty on the identity of the parameter to feed the distribution (effectively the same problem we had with the predefined functions of level 1). >> >> ? >> >> On 06/09/12 16:38, Frank T. Bergmann wrote: >>> Hello everyone, >>> >>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >>> >>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >>> >>> >>> >>> >>> </annotation> >>> >>> >>> >>> m </ci> >>> </bvar> >>> >>> s </ci> >>> </bvar> >>> m </ci> >>> </lambda> >>> </math> >>> </functionDefinition> >>> >>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >>> >>> more information and the specification here: >>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>> >>> Cheers >>> Frank >>> ____________________________________________________________ >>> To manage your sbml-discuss list subscription, visit >>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>> >>> For a web interface to the sbml-discuss mailing list, visit >>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>> >>> For questions or feedback about the sbml-discuss list, >>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>> >> >> >> -- >> Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, >> Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, >> <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere >> <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> >> >> > -- Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

Nicolas Le Novere — 2012-09-07T08:08:11-00:00

Another draft is on the way. The URI will be or the form <a href="http://www.uncertml.org/distributions/normal" target="_blank">http://www.uncertml.org/distributions/normal</a> We are working with UncertML to make sure that: * they cover everything we need If something is missing (e.g. something in iBioSim that is not covered), we must tell them, directly or through the forum <a href="http://uncertml.forum.52north.org/" target="_blank">http://uncertml.forum.52north.org/</a> * their URIs are stable Hopefully they are already. We are not their first user community * there is a versioning system in place On 07/09/12 08:59, Frank T. Bergmann wrote: > Hello, > > what an amount of replies. I've picked this one to reply to because it most specifically applies to the original post. I index thought about the distrib proposal at first. However, it was much more detailed than what i wanted to cover at first. It also would not allow me to use the new functions in for example assignment rules. > > In the end it came down to the fact that i needed to support a certain number of functions, and there was no real identifier for them (the same is true of the distrib package, in section 3.7 of the draft where the distrib:function attribute value is described does not hold much information on which values would be allowed). The draft gets away with in the examples to use for example <a href="http://identifiers.org/truncatedNormal" target="_blank">http://identifiers.org/truncatedNormal</a> or "<a href="http:///identifiers.org/Distribution/LogNormal" target="_blank">http:///identifiers.org/Distribution/LogNormal</a>", but there are no such identifiers at the moment. At the same time the distributions listed under <a href="http://www.uncertml.org/distributions" target="_blank">http://www.uncertml.org/distributions</a> did not cover the same subset as for example supported by iBioSim. > > I don't mean to imply that the wikipedia entries are what is normative in this point or that they would have to be used programmatically. It is simply a helper to identify the kind of function i want, without having to hinge on the id / name of the function definition. For me the definition attribute in the custom annotation is too be seen just as well as the namespace identifier. > > Cheers > Frank > > On Sep 6, 2012, at 6:15 PM, Nicolas Le Novère wrote: > >> Hi Frank, >> >> This is great that software start to exchange models with uncertainty. But I am puzzled by your approach of developing your own annotation rather than putting distrib constructs in the annotation element. That would be the favoured way to provide an easier path towards the package once accepted, and also minimise discrepancy between level 2 and level 3. That is at least what has been done with the layout package. >> >> In effect what you propose is very close to distrib except you use wikipedia instead of UncertML. However, since wikipedia cannot be used easily programmaticaly, there is an uncertainty on the identity of the parameter to feed the distribution (effectively the same problem we had with the predefined functions of level 1). >> >> ? >> >> On 06/09/12 16:38, Frank T. Bergmann wrote: >>> Hello everyone, >>> >>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >>> >>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >>> >>> >>> >>> >>> </annotation> >>> >>> >>> >>> m </ci> >>> </bvar> >>> >>> s </ci> >>> </bvar> >>> m </ci> >>> </lambda> >>> </math> >>> </functionDefinition> >>> >>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >>> >>> more information and the specification here: >>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>> >>> Cheers >>> Frank >>> ____________________________________________________________ >>> To manage your sbml-discuss list subscription, visit >>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>> >>> For a web interface to the sbml-discuss mailing list, visit >>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>> >>> For questions or feedback about the sbml-discuss list, >>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>> >> >> >> -- >> Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, >> Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, >> <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere >> <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> >> >> > -- Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: [sbml-discuss] Custom Annotation for drawing numbers from Distributions

fbergman — 2012-09-07T07:59:44-00:00

Hello, what an amount of replies. I've picked this one to reply to because it most specifically applies to the original post. I index thought about the distrib proposal at first. However, it was much more detailed than what i wanted to cover at first. It also would not allow me to use the new functions in for example assignment rules. In the end it came down to the fact that i needed to support a certain number of functions, and there was no real identifier for them (the same is true of the distrib package, in section 3.7 of the draft where the distrib:function attribute value is described does not hold much information on which values would be allowed). The draft gets away with in the examples to use for example <a href="http://identifiers.org/truncatedNormal" target="_blank">http://identifiers.org/truncatedNormal</a> or "<a href="http:///identifiers.org/Distribution/LogNormal" target="_blank">http:///identifiers.org/Distribution/LogNormal</a>", but there are no such identifiers at the moment. At the same time the distributions listed under <a href="http://www.uncertml.org/distributions" target="_blank">http://www.uncertml.org/distributions</a> did not cover the same subset as for example supported by iBioSim. I don't mean to imply that the wikipedia entries are what is normative in this point or that they would have to be used programmatically. It is simply a helper to identify the kind of function i want, without having to hinge on the id / name of the function definition. For me the definition attribute in the custom annotation is too be seen just as well as the namespace identifier. Cheers Frank On Sep 6, 2012, at 6:15 PM, Nicolas Le Novère wrote: > Hi Frank, > > This is great that software start to exchange models with uncertainty. But I am puzzled by your approach of developing your own annotation rather than putting distrib constructs in the annotation element. That would be the favoured way to provide an easier path towards the package once accepted, and also minimise discrepancy between level 2 and level 3. That is at least what has been done with the layout package. > > In effect what you propose is very close to distrib except you use wikipedia instead of UncertML. However, since wikipedia cannot be used easily programmaticaly, there is an uncertainty on the identity of the parameter to feed the distribution (effectively the same problem we had with the predefined functions of level 1). > > ? > > On 06/09/12 16:38, Frank T. Bergmann wrote: >> Hello everyone, >> >> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >> >> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >> >> >> >> >> </annotation> >> >> >> >> m </ci> >> </bvar> >> >> s </ci> >> </bvar> >> m </ci> >> </lambda> >> </math> >> </functionDefinition> >> >> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >> >> more information and the specification here: >> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >> >> Cheers >> Frank >> ____________________________________________________________ >> To manage your sbml-discuss list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >> >> For a web interface to the sbml-discuss mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-discuss list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >> > > > -- > Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, > Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, > <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere > <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> > > ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

fbergman — 2012-09-07T07:59:44-00:00

Hello, what an amount of replies. I've picked this one to reply to because it most specifically applies to the original post. I index thought about the distrib proposal at first. However, it was much more detailed than what i wanted to cover at first. It also would not allow me to use the new functions in for example assignment rules. In the end it came down to the fact that i needed to support a certain number of functions, and there was no real identifier for them (the same is true of the distrib package, in section 3.7 of the draft where the distrib:function attribute value is described does not hold much information on which values would be allowed). The draft gets away with in the examples to use for example <a href="http://identifiers.org/truncatedNormal" target="_blank">http://identifiers.org/truncatedNormal</a> or "<a href="http:///identifiers.org/Distribution/LogNormal" target="_blank">http:///identifiers.org/Distribution/LogNormal</a>", but there are no such identifiers at the moment. At the same time the distributions listed under <a href="http://www.uncertml.org/distributions" target="_blank">http://www.uncertml.org/distributions</a> did not cover the same subset as for example supported by iBioSim. I don't mean to imply that the wikipedia entries are what is normative in this point or that they would have to be used programmatically. It is simply a helper to identify the kind of function i want, without having to hinge on the id / name of the function definition. For me the definition attribute in the custom annotation is too be seen just as well as the namespace identifier. Cheers Frank On Sep 6, 2012, at 6:15 PM, Nicolas Le Novère wrote: > Hi Frank, > > This is great that software start to exchange models with uncertainty. But I am puzzled by your approach of developing your own annotation rather than putting distrib constructs in the annotation element. That would be the favoured way to provide an easier path towards the package once accepted, and also minimise discrepancy between level 2 and level 3. That is at least what has been done with the layout package. > > In effect what you propose is very close to distrib except you use wikipedia instead of UncertML. However, since wikipedia cannot be used easily programmaticaly, there is an uncertainty on the identity of the parameter to feed the distribution (effectively the same problem we had with the predefined functions of level 1). > > ? > > On 06/09/12 16:38, Frank T. Bergmann wrote: >> Hello everyone, >> >> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >> >> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >> >> >> >> >> </annotation> >> >> >> >> m </ci> >> </bvar> >> >> s </ci> >> </bvar> >> m </ci> >> </lambda> >> </math> >> </functionDefinition> >> >> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >> >> more information and the specification here: >> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >> >> Cheers >> Frank >> ____________________________________________________________ >> To manage your sbml-discuss list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >> >> For a web interface to the sbml-discuss mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-discuss list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >> > > > -- > Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, > Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, > <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere > <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> > > ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

hsauro — 2012-09-07T01:18:42-00:00

Assuming the algorithm is defined we should be able to store the seed because sometimes we want to recreate a stochastic simulation from a published paper in order to compute our own statistics about the simulation. Herbert On 9/6/2012 6:04 PM, Pedro Mendes wrote: > > On 09/07/2012 12:53 AM, Lucian Smith wrote: >> In SBML, you only need a pointer that says 'this is a uniform random from >> 0 (non-inclusive) to 1 (inclusive)' or the like. > Exactly! The spec should define the behaviour of the function, not its > implementation. No one in their right mind would like to define the > cos() function as a series, but that is really how computers do it > (sometimes combined with a lookup table to speed it up) > > > SED-ML would also be the place to store the random number seed, for > > what small worth that gives you (i.e. only guaranteed on the exact > > same machine that ran it the first time.) > > Seeding an RNG with a constant value is great for debugging code, > otherwise it is definitely a bug since it will make your RND sequence > completely predictable. The seed value should always be something > unpredictable. So indicating the seed in SED-ML is really worth very > little, I would even dare say: negative worth. > > -- > Pedro Mendes > Professor of Computational Systems Biology > School of Computer Science > Manchester Institute of Biotechnology > The University of Manchester, > 131 Princess St., Manchester, M1 7DN > @gepasi <a href="http://www.comp-sys-bio.org" target="_blank">http://www.comp-sys-bio.org</a> > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> -- Herbert M. Sauro, PhD Associate Professor Department of Bioengineering University of Washington William H. Foege Building Box 355061, Seattle, WA 98195-5061 206.685.2119 (tel), 206.685.3300 (shared fax) <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> New Text Book: Enzyme Kinetics for Systems Biology Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

Pedro Mendes — 2012-09-07T01:04:52-00:00

On 09/07/2012 12:53 AM, Lucian Smith wrote: > In SBML, you only need a pointer that says 'this is a uniform random from > 0 (non-inclusive) to 1 (inclusive)' or the like. Exactly! The spec should define the behaviour of the function, not its implementation. No one in their right mind would like to define the cos() function as a series, but that is really how computers do it (sometimes combined with a lookup table to speed it up) > SED-ML would also be the place to store the random number seed, for > what small worth that gives you (i.e. only guaranteed on the exact > same machine that ran it the first time.) Seeding an RNG with a constant value is great for debugging code, otherwise it is definitely a bug since it will make your RND sequence completely predictable. The seed value should always be something unpredictable. So indicating the seed in SED-ML is really worth very little, I would even dare say: negative worth. -- Pedro Mendes Professor of Computational Systems Biology School of Computer Science Manchester Institute of Biotechnology The University of Manchester, 131 Princess St., Manchester, M1 7DN @gepasi <a href="http://www.comp-sys-bio.org" target="_blank">http://www.comp-sys-bio.org</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

Lucian Smith — 2012-09-06T23:53:40-00:00

My vote would be for Frank's suggestion with UncertML URIs instead of wikipedia URIs. But even wikipedia URIs would work, and once your system was in place, it would be easy enough to change between the two. Picking a particular random number generator algorithm seems like it is out of scope for SBML, but in-scope for SED-ML, where you could use KISAO or something to have an ontology of rng algorithms that simulators could look at and say, "OK, they used the Mersenne Twister, but I can probably get by with rnd() until I implement MT properly." SED-ML would also be the place to store the random number seed, for what small worth that gives you (i.e. only guaranteed on the exact same machine that ran it the first time.) In SBML, you only need a pointer that says 'this is a uniform random from 0 (non-inclusive) to 1 (inclusive)' or the like. -Lucian * hsauro <<a href="mailto:hsauro@u.washington.edu" target="_blank">hsauro@u.washington.edu</a>> [2012-09-06 22:16] writes: > The problem is I need it this week :(. Can someone suggest a short > term fix? The best it seems so far is Frank's original suggestion > although I can't remember if he specifies the random number generator > algorithm. > > Herbert > > > On 9/6/2012 1:31 PM, Pedro Mendes wrote: > > Herbert is asking for a needed feature indeed. In COPASI we have had > > random number functions for a while, but models that include them cannot > > be exported to SBML.. > > > > However I would caution against writing out functions in the models to > > generate random numbers. Just as Herbert said, those functions that can > > be expressed in SBML are simple and they are extremely bad as random > > number generators. I don't want to get into RND genetators theory [*], > > but unfortunately the simple ones (eg rnd() functions in most languages) > > are terrible. You do not want the model to have this hard-coded in it > > because it would be nearly impossible for other software to recognize > > that the function is a generator (to be able to substitute it with a > > proper one, like the Mersenne Twister). > > > > So I would urge modellers to avoid putting in a function explicitly to > > generate random numbers. Those models will be broken forever... > > > > Pedro > > > > [*] the problem is that all RDN generators depend on generating uniform > > random numbers, but most functions are quite far from uniform. See any > > paper by Marsaglia on this (google the classic "Random numbers fall > > mainly in the planes" which is enough to illustrate this) > > > > > > On 09/06/2012 06:00 PM, hsauro wrote: > >> ok I see, thanks for the clarification. In that case one just needs to > >> include the uniform random number generator inside each user defined > >> distribution. There are however two downsides to this approach, the > >> random number generator can't be very complex (no looping in MathML, no > >> Mersenne twister?) and evaluation of the random number might be > >> inefficient at the tool end since it has to interpret the user defined > >> function. If we could have a built-in random number generator in SBML > >> just as we have sin, cos etc that would help a lot with this situation. > >> > >> Herbert Sauro > >> > >> On 9/6/2012 9:52 AM, Nicolas Le Nov?re wrote: > >>> Yes, one can do > >>> > >>> FunctionDef F1=ax+b > >>> FuntionDef F2=cy+d > >>> > >>> assignment F2(F1(z)) > >>> > >>> But one cannot do > >>> > >>> FunctionDef F1=ax+b > >>> FuntionDef F2=cF1(x)+d > >>> > >>> assignment F2(z) > >>> > >>> > >>> On 06/09/12 17:43, hsauro wrote: > >>>> I just checked the SBML spec for L2V4, couldn't figure out whether user > >>>> defined functions can call other user defined functions or not. It says > >>>> that cannot be recursive but I assume it is ok for one to call another > >>>> assuming no recursion? > >>>> > >>>> Herbert Sauro > >>>> > >>>> On 9/6/2012 9:35 AM, hsauro wrote: > >>>>> Nicolas has a point. However there is a simple way out of it. We just > >>>>> define our own uniform random number generator as a user defined > >>>>> function (we pick an algorithm we like), then declare a set of > >>>>> alternative distribution user defined functions that use the uniform > >>>>> random number generator to compute the given distribution. This saves > >>>>> work in having to upgrade to L3 and deal with packages which many of us > >>>>> don't have time to do at the moment. > >>>>> > >>>>> Herbert Sauro > >>>>> > >>>>> On 9/6/2012 9:15 AM, Nicolas Le Nov?re wrote: > >>>>>> Hi Frank, > >>>>>> > >>>>>> This is great that software start to exchange models with uncertainty. But > >>>>>> I am puzzled by your approach of developing your own annotation rather than > >>>>>> putting distrib constructs in the annotation element. That would be the > >>>>>> favoured way to provide an easier path towards the package once accepted, > >>>>>> and also minimise discrepancy between level 2 and level 3. That is at least > >>>>>> what has been done with the layout package. > >>>>>> > >>>>>> In effect what you propose is very close to distrib except you use > >>>>>> wikipedia instead of UncertML. However, since wikipedia cannot be used > >>>>>> easily programmaticaly, there is an uncertainty on the identity of the > >>>>>> parameter to feed the distribution (effectively the same problem we had > >>>>>> with the predefined functions of level 1). > >>>>>> > >>>>>> ? > >>>>>> > >>>>>> On 06/09/12 16:38, Frank T. Bergmann wrote: > >>>>>>> Hello everyone, > >>>>>>> > >>>>>>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). > >>>>>>> > >>>>>>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> </annotation> > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> m </ci> > >>>>>>> </bvar> > >>>>>>> > >>>>>>> s </ci> > >>>>>>> </bvar> > >>>>>>> m </ci> > >>>>>>> </lambda> > >>>>>>> </math> > >>>>>>> </functionDefinition> > >>>>>>> > >>>>>>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. > >>>>>>> > >>>>>>> more information and the specification here: > >>>>>>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> > >>>>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> > >>>>>>> > >>>>>>> Cheers > >>>>>>> Frank > >>>>>>> ____________________________________________________________ > >>>>>>> To manage your sbml-discuss list subscription, visit > >>>>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > >>>>>>> > >>>>>>> For a web interface to the sbml-discuss mailing list, visit > >>>>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > >>>>>>> > >>>>>>> For questions or feedback about the sbml-discuss list, > >>>>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > >>>>>>> > >> > > > > > -- > Herbert M. Sauro, PhD > Associate Professor > Department of Bioengineering > > University of Washington > William H. Foege Building > Box 355061, Seattle, WA 98195-5061 > 206.685.2119 (tel), 206.685.3300 (shared fax) > > <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> > New Text Book: Enzyme Kinetics for Systems Biology > Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> > > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

Lucian Smith — 2012-09-06T23:43:09-00:00

Exactly--in L2, this attribute was '3' by default, so all compartments knew whether they were units of volume or area or whatever. However! I came up with that list only having looked at the L2v4 spec and writing down everything that had a default. But it turns out that in L3, the attribute is optional, meaning that if it is not set, it is undefined. So it will be set if you call the 'initDefaults' function, but if you leave it unset, it will still be valid L3 SBML. -Lucian * hsauro <<a href="mailto:hsauro@u.washington.edu" target="_blank">hsauro@u.washington.edu</a>> [2012-09-06 22:16] writes: > I was trying to figure out what spatialDimensions means because I > noticed you set it to 3 in your example. Is this to maintain consistency > with any units that might be defined? > > Herbert > > On 9/6/2012 1:09 PM, Lucian Smith wrote: > > * hsauro <<a href="mailto:hsauro@u.washington.edu" target="_blank">hsauro@u.washington.edu</a>> [2012-09-06 20:40] writes: > >> Sounds like Chris Myers has done something with packages, maybe you > >> could pick he brains for a suggested tutorial. > >> > >> I am sure there used to be some kind of document on how to work with > >> packages but I can't find it anymore. > > Hmm, I think I remember this too. I'll check with Frank; I think he wrote > > it. > > > >> For moving from L2 to L3, any possible gotchas we should be aware of? > > The biggest thing for myself was the fact that you must now set explicitly > > what used to be there by default, or the resulting model won't be valid > > (or, if you create the element but don't set the needs-a-default > > attributes, libsbml won't let you add it to the model.) > > > > This includes: > > > > Unit: exponent, scale, multiplier > > Compartment: spatialDimensions, constant > > Species: hasOnlySubstanceUnits, boundaryCondition, constant > > Parameter: constant > > Reaction: reversible, fast > > SpeciesReference: stoichiometry > > Event: useValuesFromTriggerTime > > > > You can either set these explicitly in your code when you create those > > objects (which would be my recommendation) or you can use the function > > 'initDefaults', which will (for example), if you call it on a Species > > object, set those three attributes to their default L2v4 values. (This > > includes units--'initDefaults' called on a Compartment will explicitly set > > the units to 'litre', like they were assumed to be for L2.) > > > > -Lucian > > ____________________________________________________________ > > To manage your sbml-discuss list subscription, visit > > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > > > For a web interface to the sbml-discuss mailing list, visit > > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > > > For questions or feedback about the sbml-discuss list, > > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > > -- > Herbert M. Sauro, PhD > Associate Professor > Department of Bioengineering > > University of Washington > William H. Foege Building > Box 355061, Seattle, WA 98195-5061 > 206.685.2119 (tel), 206.685.3300 (shared fax) > > <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> > New Text Book: Enzyme Kinetics for Systems Biology > Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> > > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Model_add functions not updating model properly.

Lucian Smith — 2012-09-06T23:25:22-00:00

* mihajlv <<a href="mailto:mihajlv@gmail.com" target="_blank">mihajlv@gmail.com</a>> [2012-09-06 23:00] writes: > > I am doing the following for various entities but the model doesn't get updated on any of them: > > UnitDefinition_t * ud = UnitDefinition_create(3,1); > UnitDefinition_setId(ud, "some_unit"); > unit_definition_add_units(ud, UNIT_KIND_SECOND); > Model_addUnitDefinition(model, ud); My guess is that this is the function that isn't working--all these functions now return a value, so you can check to see what the return value is--anything other than '0' (LIBSBML_OPERATION_SUCCESS) means that something has gone wrong. In this case, I believe the problem is that the 'Unit' element is invalid SBML Level 3: you must explicitly define the exponent, scale, and multiplier for all of them. The various 'add' functions refuse to let you create invalid SBML. > However, If I create the unit definition with > > UnitDefinition_t * ud = Model_createUnitDefinition(model) > > and the unit with > > Unit_t * u = UnitDefinition_createUnit(ud); > > then the code works. Am I missing something? Thanks for the help. In contrast to the 'add' functions, the 'create' functions do indeed let you create invalid SBML in that they let you create objects with undefined required attributes. If you were to try to write out this model with the 'strict' flag on, it would refuse. Amusingly, this is exactly the 'gotcha' possible with changing your code from supporting L2 to supporting L3 that Chris and I were talking about with Herbet today on sbml-discuss! -Lucian ____________________________________________________________ To manage your sbml-interoperability list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-interoperability</a> For a web interface to the sbml-interoperability mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-interoperability list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Model_add functions not updating model properly.

mihajlv — 2012-09-06T21:54:58-00:00

I am doing the following for various entities but the model doesn't get updated on any of them: UnitDefinition_t * ud = UnitDefinition_create(3,1); UnitDefinition_setId(ud, "some_unit"); unit_definition_add_units(ud, UNIT_KIND_SECOND); Model_addUnitDefinition(model, ud); Model_addUnitDefinition produces LIBSBML_INVALID_OBJECT However, If I create the unit definition with UnitDefinition_t * ud = Model_createUnitDefinition(model) and the unit with Unit_t * u = UnitDefinition_createUnit(ud); then the code works. Am I missing something? Thanks for the help.

Re: Custom Annotation for drawing numbers from Distributions

hsauro — 2012-09-06T21:20:30-00:00

Thanks for that, your slides are useful. Looks like more of a long term upgrade rather than a quick fix. For now I'm going to go with Frank's interim solution and then eventually migrate to L3 Distrib. Herbert On 9/6/2012 1:02 PM, Chris J. Myers wrote: > Hi, > > Here are a couple of presentations that I've given at COMBINE on L3 support and packages, respectively. There are only a few issues you need to worry about for L3V1 which are described on slides 4 and 5 of the first talk. The only major thing is there are some added features with events which is the majority of this presentation. You may recall that these are the new features I requested at the Hackathon in Seattle. > > As for packages, if you check out slide 30 of the 2nd talk, you can see how you can immediately begin importing models with hierarchy. Basically, you read it and flatten it. Now, you actually want to use packages, the key thing is the idea that packages are a plugin. Again, on slide 30, you see the line that get's the plugin for comp. This is an object which contains all the model additions related to comp. For example, to get a species "foo" in Java, you say (assuming the first 3 lines of the code on slide 30): > > document.getModel().getSpecies("foo"); > > To get a port "foo", you would do: > > compModel.getPort("foo"); > > Once you get your head around this fact, the rest is pretty straightforward. > > Feel free to ask me further questions. It is really not that bad, and SBML L3 with packages is so much more powerful L2. This is also the right time to get into it as several packages now have libsbml support and are nearing final approval. > > Cheers, > > Chris > > > > > On Sep 6, 2012, at 1:02 PM, hsauro wrote: > >> Sounds like Chris Myers has done something with packages, maybe you >> could pick he brains for a suggested tutorial. >> >> I am sure there used to be some kind of document on how to work with >> packages but I can't find it anymore. >> >> For moving from L2 to L3, any possible gotchas we should be aware of? >> >> H >> >> On 9/6/2012 11:28 AM, Lucian Smith wrote: >>> Yes, the idea is that you would use libSBML. Different packages have >>> different levels of libSBML support--for hierarchy, I tried to include a >>> lot of example code so that getting started would be easier, and I believe >>> a lot of the other packages have a good amount of example code as well. >>> >>> However, now that libSBML package support has matured enough, you're right >>> that some sort of tutorial on how to code for packages in general would be >>> great, along with more specific tutorials on coding for particular >>> packages. Time is always the issue, but I bet getting something together >>> for HARMONY would be possible, if there was enough interest. >>> >>> -Lucian >>> >>> * hsauro <<a href="mailto:hsauro@u.washington.edu" target="_blank">hsauro@u.washington.edu</a>> [2012-09-06 18:48] writes: >>>> Its supporting the packages that scares me in terms of effort expended. >>>> I tried looking up documentation on the sbml.org site this morning on >>>> how to work with packages but I couldn't find anything other than a list >>>> of packages in progress. I thought there was some docs somewhere. I >>>> looked at the distrib package docs and found this little handy snip bit >>>> from one of the examples: >>>> >>>> >>>> >>>> >>>> >>>> 3.14</un:logScale> >>>> 3.14</un:shape> >>>> </un:LogNormalDistribution> >>>> </distrib:distribution> >>>> >>>> >>>> mu</ci></bvar> >>>> sigma</ci></bvar> >>>> mu</ci> >>>> </lambda> >>>> </math> >>>> </functionDefinition> >>>> >>>> What do I actually do with it? I want to avoid working through the XML >>>> myself. Can I use libSBML to pull the information out without too much >>>> effort? If its too hard to pull the info out, very few people are likely >>>> use it other than those with the necessary human resources. >>>> >>>> Herbert >>>> >>>> >>>> On 9/6/2012 10:10 AM, Chris J. Myers wrote: >>>>> That is direction we are heading with distrib which I think should be ready very soon. Stuart and Sarah can comment on this. >>>>> >>>>> As for not moving to L3, this is not difficult as it is very similar to L2. You only need to deal with those packages you find useful. Switching to L3 for us was trivial. Making use of packages took some work, but it was really worth it as it allowed us to replace much of the specialized stuff we were doing in our tool that made our models non-exchangeable. Hierarchical modeling which is now complete and should have final approval this month should be incentive enough to move to L3, if you ask me. >>>>> >>>>> Chris >>>>> >>>>> On Sep 6, 2012, at 11:00 AM, hsauro wrote: >>>>> >>>>>> ok I see, thanks for the clarification. In that case one just needs to >>>>>> include the uniform random number generator inside each user defined >>>>>> distribution. There are however two downsides to this approach, the >>>>>> random number generator can't be very complex (no looping in MathML, no >>>>>> Mersenne twister?) and evaluation of the random number might be >>>>>> inefficient at the tool end since it has to interpret the user defined >>>>>> function. If we could have a built-in random number generator in SBML >>>>>> just as we have sin, cos etc that would help a lot with this situation. >>>>>> >>>>>> Herbert Sauro >>>>>> >>>>>> On 9/6/2012 9:52 AM, Nicolas Le Nov?re wrote: >>>>>>> Yes, one can do >>>>>>> >>>>>>> FunctionDef F1=ax+b >>>>>>> FuntionDef F2=cy+d >>>>>>> >>>>>>> assignment F2(F1(z)) >>>>>>> >>>>>>> But one cannot do >>>>>>> >>>>>>> FunctionDef F1=ax+b >>>>>>> FuntionDef F2=cF1(x)+d >>>>>>> >>>>>>> assignment F2(z) >>>>>>> >>>>>>> >>>>>>> On 06/09/12 17:43, hsauro wrote: >>>>>>>> I just checked the SBML spec for L2V4, couldn't figure out whether user >>>>>>>> defined functions can call other user defined functions or not. It says >>>>>>>> that cannot be recursive but I assume it is ok for one to call another >>>>>>>> assuming no recursion? >>>>>>>> >>>>>>>> Herbert Sauro >>>>>>>> >>>>>>>> On 9/6/2012 9:35 AM, hsauro wrote: >>>>>>>>> Nicolas has a point. However there is a simple way out of it. We just >>>>>>>>> define our own uniform random number generator as a user defined >>>>>>>>> function (we pick an algorithm we like), then declare a set of >>>>>>>>> alternative distribution user defined functions that use the uniform >>>>>>>>> random number generator to compute the given distribution. This saves >>>>>>>>> work in having to upgrade to L3 and deal with packages which many of us >>>>>>>>> don't have time to do at the moment. >>>>>>>>> >>>>>>>>> Herbert Sauro >>>>>>>>> >>>>>>>>> On 9/6/2012 9:15 AM, Nicolas Le Nov?re wrote: >>>>>>>>>> Hi Frank, >>>>>>>>>> >>>>>>>>>> This is great that software start to exchange models with uncertainty. But >>>>>>>>>> I am puzzled by your approach of developing your own annotation rather than >>>>>>>>>> putting distrib constructs in the annotation element. That would be the >>>>>>>>>> favoured way to provide an easier path towards the package once accepted, >>>>>>>>>> and also minimise discrepancy between level 2 and level 3. That is at least >>>>>>>>>> what has been done with the layout package. >>>>>>>>>> >>>>>>>>>> In effect what you propose is very close to distrib except you use >>>>>>>>>> wikipedia instead of UncertML. However, since wikipedia cannot be used >>>>>>>>>> easily programmaticaly, there is an uncertainty on the identity of the >>>>>>>>>> parameter to feed the distribution (effectively the same problem we had >>>>>>>>>> with the predefined functions of level 1). >>>>>>>>>> >>>>>>>>>> ? >>>>>>>>>> >>>>>>>>>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>>>>>>>>> Hello everyone, >>>>>>>>>>> >>>>>>>>>>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >>>>>>>>>>> >>>>>>>>>>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> </annotation> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> m </ci> >>>>>>>>>>> </bvar> >>>>>>>>>>> >>>>>>>>>>> s </ci> >>>>>>>>>>> </bvar> >>>>>>>>>>> m </ci> >>>>>>>>>>> </lambda> >>>>>>>>>>> </math> >>>>>>>>>>> </functionDefinition> >>>>>>>>>>> >>>>>>>>>>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >>>>>>>>>>> >>>>>>>>>>> more information and the specification here: >>>>>>>>>>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>>>>>>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>>>>>>>>> >>>>>>>>>>> Cheers >>>>>>>>>>> Frank >>>>>>>>>>> ____________________________________________________________ >>>>>>>>>>> To manage your sbml-discuss list subscription, visit >>>>>>>>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>>>>>>>> >>>>>>>>>>> For a web interface to the sbml-discuss mailing list, visit >>>>>>>>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>>>>>>>> >>>>>>>>>>> For questions or feedback about the sbml-discuss list, >>>>>>>>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>>>>>>>> >>>>>> -- >>>>>> Herbert M. Sauro, PhD >>>>>> Associate Professor >>>>>> Department of Bioengineering >>>>>> >>>>>> University of Washington >>>>>> William H. Foege Building >>>>>> Box 355061, Seattle, WA 98195-5061 >>>>>> 206.685.2119 (tel), 206.685.3300 (shared fax) >>>>>> >>>>>> <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> >>>>>> New Text Book: Enzyme Kinetics for Systems Biology >>>>>> Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> >>>>>> >>>>>> ____________________________________________________________ >>>>>> To manage your sbml-discuss list subscription, visit >>>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>>> >>>>>> For a web interface to the sbml-discuss mailing list, visit >>>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>>> >>>>>> For questions or feedback about the sbml-discuss list, >>>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>> ____________________________________________________________ >>>>> To manage your sbml-discuss list subscription, visit >>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>> >>>>> For a web interface to the sbml-discuss mailing list, visit >>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>> >>>>> For questions or feedback about the sbml-discuss list, >>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>> -- >>>> Herbert M. Sauro, PhD >>>> Associate Professor >>>> Department of Bioengineering >>>> >>>> University of Washington >>>> William H. Foege Building >>>> Box 355061, Seattle, WA 98195-5061 >>>> 206.685.2119 (tel), 206.685.3300 (shared fax) >>>> >>>> <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> >>>> New Text Book: Enzyme Kinetics for Systems Biology >>>> Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> >>>> >>>> ____________________________________________________________ >>>> To manage your sbml-discuss list subscription, visit >>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>> >>>> For a web interface to the sbml-discuss mailing list, visit >>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>> >>>> For questions or feedback about the sbml-discuss list, >>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>> ____________________________________________________________ >>> To manage your sbml-discuss list subscription, visit >>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>> >>> For a web interface to the sbml-discuss mailing list, visit >>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>> >>> For questions or feedback about the sbml-discuss list, >>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >> >> -- >> Herbert M. Sauro, PhD >> Associate Professor >> Department of Bioengineering >> >> University of Washington >> William H. Foege Building >> Box 355061, Seattle, WA 98195-5061 >> 206.685.2119 (tel), 206.685.3300 (shared fax) >> >> <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> >> New Text Book: Enzyme Kinetics for Systems Biology >> Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> >> >> ____________________________________________________________ >> To manage your sbml-discuss list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >> >> For a web interface to the sbml-discuss mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-discuss list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> -- Herbert M. Sauro, PhD Associate Professor Department of Bioengineering University of Washington William H. Foege Building Box 355061, Seattle, WA 98195-5061 206.685.2119 (tel), 206.685.3300 (shared fax) <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> New Text Book: Enzyme Kinetics for Systems Biology Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

hsauro — 2012-09-06T21:11:40-00:00

The problem is I need it this week :(. Can someone suggest a short term fix? The best it seems so far is Frank's original suggestion although I can't remember if he specifies the random number generator algorithm. Herbert On 9/6/2012 1:31 PM, Pedro Mendes wrote: > Herbert is asking for a needed feature indeed. In COPASI we have had > random number functions for a while, but models that include them cannot > be exported to SBML.. > > However I would caution against writing out functions in the models to > generate random numbers. Just as Herbert said, those functions that can > be expressed in SBML are simple and they are extremely bad as random > number generators. I don't want to get into RND genetators theory [*], > but unfortunately the simple ones (eg rnd() functions in most languages) > are terrible. You do not want the model to have this hard-coded in it > because it would be nearly impossible for other software to recognize > that the function is a generator (to be able to substitute it with a > proper one, like the Mersenne Twister). > > So I would urge modellers to avoid putting in a function explicitly to > generate random numbers. Those models will be broken forever... > > Pedro > > [*] the problem is that all RDN generators depend on generating uniform > random numbers, but most functions are quite far from uniform. See any > paper by Marsaglia on this (google the classic "Random numbers fall > mainly in the planes" which is enough to illustrate this) > > > On 09/06/2012 06:00 PM, hsauro wrote: >> ok I see, thanks for the clarification. In that case one just needs to >> include the uniform random number generator inside each user defined >> distribution. There are however two downsides to this approach, the >> random number generator can't be very complex (no looping in MathML, no >> Mersenne twister?) and evaluation of the random number might be >> inefficient at the tool end since it has to interpret the user defined >> function. If we could have a built-in random number generator in SBML >> just as we have sin, cos etc that would help a lot with this situation. >> >> Herbert Sauro >> >> On 9/6/2012 9:52 AM, Nicolas Le NovÃ¨re wrote: >>> Yes, one can do >>> >>> FunctionDef F1=ax+b >>> FuntionDef F2=cy+d >>> >>> assignment F2(F1(z)) >>> >>> But one cannot do >>> >>> FunctionDef F1=ax+b >>> FuntionDef F2=cF1(x)+d >>> >>> assignment F2(z) >>> >>> >>> On 06/09/12 17:43, hsauro wrote: >>>> I just checked the SBML spec for L2V4, couldn't figure out whether user >>>> defined functions can call other user defined functions or not. It says >>>> that cannot be recursive but I assume it is ok for one to call another >>>> assuming no recursion? >>>> >>>> Herbert Sauro >>>> >>>> On 9/6/2012 9:35 AM, hsauro wrote: >>>>> Nicolas has a point. However there is a simple way out of it. We just >>>>> define our own uniform random number generator as a user defined >>>>> function (we pick an algorithm we like), then declare a set of >>>>> alternative distribution user defined functions that use the uniform >>>>> random number generator to compute the given distribution. This saves >>>>> work in having to upgrade to L3 and deal with packages which many of us >>>>> don't have time to do at the moment. >>>>> >>>>> Herbert Sauro >>>>> >>>>> On 9/6/2012 9:15 AM, Nicolas Le NovÃ¨re wrote: >>>>>> Hi Frank, >>>>>> >>>>>> This is great that software start to exchange models with uncertainty. But >>>>>> I am puzzled by your approach of developing your own annotation rather than >>>>>> putting distrib constructs in the annotation element. That would be the >>>>>> favoured way to provide an easier path towards the package once accepted, >>>>>> and also minimise discrepancy between level 2 and level 3. That is at least >>>>>> what has been done with the layout package. >>>>>> >>>>>> In effect what you propose is very close to distrib except you use >>>>>> wikipedia instead of UncertML. However, since wikipedia cannot be used >>>>>> easily programmaticaly, there is an uncertainty on the identity of the >>>>>> parameter to feed the distribution (effectively the same problem we had >>>>>> with the predefined functions of level 1). >>>>>> >>>>>> ? >>>>>> >>>>>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>>>>> Hello everyone, >>>>>>> >>>>>>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >>>>>>> >>>>>>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> </annotation> >>>>>>> >>>>>>> >>>>>>> >>>>>>> m </ci> >>>>>>> </bvar> >>>>>>> >>>>>>> s </ci> >>>>>>> </bvar> >>>>>>> m </ci> >>>>>>> </lambda> >>>>>>> </math> >>>>>>> </functionDefinition> >>>>>>> >>>>>>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >>>>>>> >>>>>>> more information and the specification here: >>>>>>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>>>>> >>>>>>> Cheers >>>>>>> Frank >>>>>>> ____________________________________________________________ >>>>>>> To manage your sbml-discuss list subscription, visit >>>>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>>>> >>>>>>> For a web interface to the sbml-discuss mailing list, visit >>>>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>>>> >>>>>>> For questions or feedback about the sbml-discuss list, >>>>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>>>> >> > -- Herbert M. Sauro, PhD Associate Professor Department of Bioengineering University of Washington William H. Foege Building Box 355061, Seattle, WA 98195-5061 206.685.2119 (tel), 206.685.3300 (shared fax) <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> New Text Book: Enzyme Kinetics for Systems Biology Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

hsauro — 2012-09-06T21:07:19-00:00

I was trying to figure out what spatialDimensions means because I noticed you set it to 3 in your example. Is this to maintain consistency with any units that might be defined? Herbert On 9/6/2012 1:09 PM, Lucian Smith wrote: > * hsauro <<a href="mailto:hsauro@u.washington.edu" target="_blank">hsauro@u.washington.edu</a>> [2012-09-06 20:40] writes: >> Sounds like Chris Myers has done something with packages, maybe you >> could pick he brains for a suggested tutorial. >> >> I am sure there used to be some kind of document on how to work with >> packages but I can't find it anymore. > Hmm, I think I remember this too. I'll check with Frank; I think he wrote > it. > >> For moving from L2 to L3, any possible gotchas we should be aware of? > The biggest thing for myself was the fact that you must now set explicitly > what used to be there by default, or the resulting model won't be valid > (or, if you create the element but don't set the needs-a-default > attributes, libsbml won't let you add it to the model.) > > This includes: > > Unit: exponent, scale, multiplier > Compartment: spatialDimensions, constant > Species: hasOnlySubstanceUnits, boundaryCondition, constant > Parameter: constant > Reaction: reversible, fast > SpeciesReference: stoichiometry > Event: useValuesFromTriggerTime > > You can either set these explicitly in your code when you create those > objects (which would be my recommendation) or you can use the function > 'initDefaults', which will (for example), if you call it on a Species > object, set those three attributes to their default L2v4 values. (This > includes units--'initDefaults' called on a Compartment will explicitly set > the units to 'litre', like they were assumed to be for L2.) > > -Lucian > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> -- Herbert M. Sauro, PhD Associate Professor Department of Bioengineering University of Washington William H. Foege Building Box 355061, Seattle, WA 98195-5061 206.685.2119 (tel), 206.685.3300 (shared fax) <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> New Text Book: Enzyme Kinetics for Systems Biology Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

Pedro Mendes — 2012-09-06T20:31:50-00:00

Herbert is asking for a needed feature indeed. In COPASI we have had random number functions for a while, but models that include them cannot be exported to SBML.. However I would caution against writing out functions in the models to generate random numbers. Just as Herbert said, those functions that can be expressed in SBML are simple and they are extremely bad as random number generators. I don't want to get into RND genetators theory [*], but unfortunately the simple ones (eg rnd() functions in most languages) are terrible. You do not want the model to have this hard-coded in it because it would be nearly impossible for other software to recognize that the function is a generator (to be able to substitute it with a proper one, like the Mersenne Twister). So I would urge modellers to avoid putting in a function explicitly to generate random numbers. Those models will be broken forever... Pedro [*] the problem is that all RDN generators depend on generating uniform random numbers, but most functions are quite far from uniform. See any paper by Marsaglia on this (google the classic "Random numbers fall mainly in the planes" which is enough to illustrate this) On 09/06/2012 06:00 PM, hsauro wrote: > ok I see, thanks for the clarification. In that case one just needs to > include the uniform random number generator inside each user defined > distribution. There are however two downsides to this approach, the > random number generator can't be very complex (no looping in MathML, no > Mersenne twister?) and evaluation of the random number might be > inefficient at the tool end since it has to interpret the user defined > function. If we could have a built-in random number generator in SBML > just as we have sin, cos etc that would help a lot with this situation. > > Herbert Sauro > > On 9/6/2012 9:52 AM, Nicolas Le NovÃ¨re wrote: >> Yes, one can do >> >> FunctionDef F1=ax+b >> FuntionDef F2=cy+d >> >> assignment F2(F1(z)) >> >> But one cannot do >> >> FunctionDef F1=ax+b >> FuntionDef F2=cF1(x)+d >> >> assignment F2(z) >> >> >> On 06/09/12 17:43, hsauro wrote: >>> I just checked the SBML spec for L2V4, couldn't figure out whether user >>> defined functions can call other user defined functions or not. It says >>> that cannot be recursive but I assume it is ok for one to call another >>> assuming no recursion? >>> >>> Herbert Sauro >>> >>> On 9/6/2012 9:35 AM, hsauro wrote: >>>> Nicolas has a point. However there is a simple way out of it. We just >>>> define our own uniform random number generator as a user defined >>>> function (we pick an algorithm we like), then declare a set of >>>> alternative distribution user defined functions that use the uniform >>>> random number generator to compute the given distribution. This saves >>>> work in having to upgrade to L3 and deal with packages which many of us >>>> don't have time to do at the moment. >>>> >>>> Herbert Sauro >>>> >>>> On 9/6/2012 9:15 AM, Nicolas Le NovÃ¨re wrote: >>>>> Hi Frank, >>>>> >>>>> This is great that software start to exchange models with uncertainty. But >>>>> I am puzzled by your approach of developing your own annotation rather than >>>>> putting distrib constructs in the annotation element. That would be the >>>>> favoured way to provide an easier path towards the package once accepted, >>>>> and also minimise discrepancy between level 2 and level 3. That is at least >>>>> what has been done with the layout package. >>>>> >>>>> In effect what you propose is very close to distrib except you use >>>>> wikipedia instead of UncertML. However, since wikipedia cannot be used >>>>> easily programmaticaly, there is an uncertainty on the identity of the >>>>> parameter to feed the distribution (effectively the same problem we had >>>>> with the predefined functions of level 1). >>>>> >>>>> ? >>>>> >>>>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>>>> Hello everyone, >>>>>> >>>>>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >>>>>> >>>>>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> </annotation> >>>>>> >>>>>> >>>>>> >>>>>> m </ci> >>>>>> </bvar> >>>>>> >>>>>> s </ci> >>>>>> </bvar> >>>>>> m </ci> >>>>>> </lambda> >>>>>> </math> >>>>>> </functionDefinition> >>>>>> >>>>>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >>>>>> >>>>>> more information and the specification here: >>>>>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>>>> >>>>>> Cheers >>>>>> Frank >>>>>> ____________________________________________________________ >>>>>> To manage your sbml-discuss list subscription, visit >>>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>>> >>>>>> For a web interface to the sbml-discuss mailing list, visit >>>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>>> >>>>>> For questions or feedback about the sbml-discuss list, >>>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>>> >>> >> > > -- Pedro Mendes Professor of Computational Systems Biology School of Computer Science Manchester Institute of Biotechnology The University of Manchester, 131 Princess St., Manchester, M1 7DN @gepasi <a href="http://www.comp-sys-bio.org" target="_blank">http://www.comp-sys-bio.org</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

myers — 2012-09-06T20:25:15-00:00

One more thing I remembered (not sure if this is an L3 thing or not), but I had to be careful about the add versus create object methods. When you create an object, it inherits the level and version of the document you create it in. However, when you use a constructor instead and then try to add, you need to be sure that the constructor is told the correct level and version, so when you add it to the document, there is no problem. Speaking of which, the add methods now return failure when there is a problem. Since they previously did not have return methods, I was not checking these. One of the main reasons this fails is you have not set the defaults as Lucian pointed out. This can definitely bite you if you are not careful. Chris On Sep 6, 2012, at 2:09 PM, Lucian Smith wrote: > * hsauro <<a href="mailto:hsauro@u.washington.edu" target="_blank">hsauro@u.washington.edu</a>> [2012-09-06 20:40] writes: >> Sounds like Chris Myers has done something with packages, maybe you >> could pick he brains for a suggested tutorial. >> >> I am sure there used to be some kind of document on how to work with >> packages but I can't find it anymore. > > Hmm, I think I remember this too. I'll check with Frank; I think he wrote > it. > >> For moving from L2 to L3, any possible gotchas we should be aware of? > > The biggest thing for myself was the fact that you must now set explicitly > what used to be there by default, or the resulting model won't be valid > (or, if you create the element but don't set the needs-a-default > attributes, libsbml won't let you add it to the model.) > > This includes: > > Unit: exponent, scale, multiplier > Compartment: spatialDimensions, constant > Species: hasOnlySubstanceUnits, boundaryCondition, constant > Parameter: constant > Reaction: reversible, fast > SpeciesReference: stoichiometry > Event: useValuesFromTriggerTime > > You can either set these explicitly in your code when you create those > objects (which would be my recommendation) or you can use the function > 'initDefaults', which will (for example), if you call it on a Species > object, set those three attributes to their default L2v4 values. (This > includes units--'initDefaults' called on a Compartment will explicitly set > the units to 'litre', like they were assumed to be for L2.) > > -Lucian > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

Lucian Smith — 2012-09-06T20:09:29-00:00

* hsauro <<a href="mailto:hsauro@u.washington.edu" target="_blank">hsauro@u.washington.edu</a>> [2012-09-06 20:40] writes: > Sounds like Chris Myers has done something with packages, maybe you > could pick he brains for a suggested tutorial. > > I am sure there used to be some kind of document on how to work with > packages but I can't find it anymore. Hmm, I think I remember this too. I'll check with Frank; I think he wrote it. > For moving from L2 to L3, any possible gotchas we should be aware of? The biggest thing for myself was the fact that you must now set explicitly what used to be there by default, or the resulting model won't be valid (or, if you create the element but don't set the needs-a-default attributes, libsbml won't let you add it to the model.) This includes: Unit: exponent, scale, multiplier Compartment: spatialDimensions, constant Species: hasOnlySubstanceUnits, boundaryCondition, constant Parameter: constant Reaction: reversible, fast SpeciesReference: stoichiometry Event: useValuesFromTriggerTime You can either set these explicitly in your code when you create those objects (which would be my recommendation) or you can use the function 'initDefaults', which will (for example), if you call it on a Species object, set those three attributes to their default L2v4 values. (This includes units--'initDefaults' called on a Compartment will explicitly set the units to 'litre', like they were assumed to be for L2.) -Lucian ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

myers — 2012-09-06T20:02:56-00:00

Hi, Here are a couple of presentations that I've given at COMBINE on L3 support and packages, respectively. There are only a few issues you need to worry about for L3V1 which are described on slides 4 and 5 of the first talk. The only major thing is there are some added features with events which is the majority of this presentation. You may recall that these are the new features I requested at the Hackathon in Seattle. As for packages, if you check out slide 30 of the 2nd talk, you can see how you can immediately begin importing models with hierarchy. Basically, you read it and flatten it. Now, you actually want to use packages, the key thing is the idea that packages are a plugin. Again, on slide 30, you see the line that get's the plugin for comp. This is an object which contains all the model additions related to comp. For example, to get a species "foo" in Java, you say (assuming the first 3 lines of the code on slide 30): document.getModel().getSpecies("foo"); To get a port "foo", you would do: compModel.getPort("foo"); Once you get your head around this fact, the rest is pretty straightforward. Feel free to ask me further questions. It is really not that bad, and SBML L3 with packages is so much more powerful L2. This is also the right time to get into it as several packages now have libsbml support and are nearing final approval. Cheers, Chris

Re: Custom Annotation for drawing numbers from Distributions

hsauro — 2012-09-06T19:02:12-00:00

Sounds like Chris Myers has done something with packages, maybe you could pick he brains for a suggested tutorial. I am sure there used to be some kind of document on how to work with packages but I can't find it anymore. For moving from L2 to L3, any possible gotchas we should be aware of? H On 9/6/2012 11:28 AM, Lucian Smith wrote: > Yes, the idea is that you would use libSBML. Different packages have > different levels of libSBML support--for hierarchy, I tried to include a > lot of example code so that getting started would be easier, and I believe > a lot of the other packages have a good amount of example code as well. > > However, now that libSBML package support has matured enough, you're right > that some sort of tutorial on how to code for packages in general would be > great, along with more specific tutorials on coding for particular > packages. Time is always the issue, but I bet getting something together > for HARMONY would be possible, if there was enough interest. > > -Lucian > > * hsauro <<a href="mailto:hsauro@u.washington.edu" target="_blank">hsauro@u.washington.edu</a>> [2012-09-06 18:48] writes: >> Its supporting the packages that scares me in terms of effort expended. >> I tried looking up documentation on the sbml.org site this morning on >> how to work with packages but I couldn't find anything other than a list >> of packages in progress. I thought there was some docs somewhere. I >> looked at the distrib package docs and found this little handy snip bit >> from one of the examples: >> >> >> >> >> >> 3.14</un:logScale> >> 3.14</un:shape> >> </un:LogNormalDistribution> >> </distrib:distribution> >> >> >> mu</ci></bvar> >> sigma</ci></bvar> >> mu</ci> >> </lambda> >> </math> >> </functionDefinition> >> >> What do I actually do with it? I want to avoid working through the XML >> myself. Can I use libSBML to pull the information out without too much >> effort? If its too hard to pull the info out, very few people are likely >> use it other than those with the necessary human resources. >> >> Herbert >> >> >> On 9/6/2012 10:10 AM, Chris J. Myers wrote: >>> That is direction we are heading with distrib which I think should be ready very soon. Stuart and Sarah can comment on this. >>> >>> As for not moving to L3, this is not difficult as it is very similar to L2. You only need to deal with those packages you find useful. Switching to L3 for us was trivial. Making use of packages took some work, but it was really worth it as it allowed us to replace much of the specialized stuff we were doing in our tool that made our models non-exchangeable. Hierarchical modeling which is now complete and should have final approval this month should be incentive enough to move to L3, if you ask me. >>> >>> Chris >>> >>> On Sep 6, 2012, at 11:00 AM, hsauro wrote: >>> >>>> ok I see, thanks for the clarification. In that case one just needs to >>>> include the uniform random number generator inside each user defined >>>> distribution. There are however two downsides to this approach, the >>>> random number generator can't be very complex (no looping in MathML, no >>>> Mersenne twister?) and evaluation of the random number might be >>>> inefficient at the tool end since it has to interpret the user defined >>>> function. If we could have a built-in random number generator in SBML >>>> just as we have sin, cos etc that would help a lot with this situation. >>>> >>>> Herbert Sauro >>>> >>>> On 9/6/2012 9:52 AM, Nicolas Le Nov?re wrote: >>>>> Yes, one can do >>>>> >>>>> FunctionDef F1=ax+b >>>>> FuntionDef F2=cy+d >>>>> >>>>> assignment F2(F1(z)) >>>>> >>>>> But one cannot do >>>>> >>>>> FunctionDef F1=ax+b >>>>> FuntionDef F2=cF1(x)+d >>>>> >>>>> assignment F2(z) >>>>> >>>>> >>>>> On 06/09/12 17:43, hsauro wrote: >>>>>> I just checked the SBML spec for L2V4, couldn't figure out whether user >>>>>> defined functions can call other user defined functions or not. It says >>>>>> that cannot be recursive but I assume it is ok for one to call another >>>>>> assuming no recursion? >>>>>> >>>>>> Herbert Sauro >>>>>> >>>>>> On 9/6/2012 9:35 AM, hsauro wrote: >>>>>>> Nicolas has a point. However there is a simple way out of it. We just >>>>>>> define our own uniform random number generator as a user defined >>>>>>> function (we pick an algorithm we like), then declare a set of >>>>>>> alternative distribution user defined functions that use the uniform >>>>>>> random number generator to compute the given distribution. This saves >>>>>>> work in having to upgrade to L3 and deal with packages which many of us >>>>>>> don't have time to do at the moment. >>>>>>> >>>>>>> Herbert Sauro >>>>>>> >>>>>>> On 9/6/2012 9:15 AM, Nicolas Le Nov?re wrote: >>>>>>>> Hi Frank, >>>>>>>> >>>>>>>> This is great that software start to exchange models with uncertainty. But >>>>>>>> I am puzzled by your approach of developing your own annotation rather than >>>>>>>> putting distrib constructs in the annotation element. That would be the >>>>>>>> favoured way to provide an easier path towards the package once accepted, >>>>>>>> and also minimise discrepancy between level 2 and level 3. That is at least >>>>>>>> what has been done with the layout package. >>>>>>>> >>>>>>>> In effect what you propose is very close to distrib except you use >>>>>>>> wikipedia instead of UncertML. However, since wikipedia cannot be used >>>>>>>> easily programmaticaly, there is an uncertainty on the identity of the >>>>>>>> parameter to feed the distribution (effectively the same problem we had >>>>>>>> with the predefined functions of level 1). >>>>>>>> >>>>>>>> ? >>>>>>>> >>>>>>>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>>>>>>> Hello everyone, >>>>>>>>> >>>>>>>>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >>>>>>>>> >>>>>>>>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> </annotation> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> m </ci> >>>>>>>>> </bvar> >>>>>>>>> >>>>>>>>> s </ci> >>>>>>>>> </bvar> >>>>>>>>> m </ci> >>>>>>>>> </lambda> >>>>>>>>> </math> >>>>>>>>> </functionDefinition> >>>>>>>>> >>>>>>>>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >>>>>>>>> >>>>>>>>> more information and the specification here: >>>>>>>>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>>>>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>>>>>>> >>>>>>>>> Cheers >>>>>>>>> Frank >>>>>>>>> ____________________________________________________________ >>>>>>>>> To manage your sbml-discuss list subscription, visit >>>>>>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>>>>>> >>>>>>>>> For a web interface to the sbml-discuss mailing list, visit >>>>>>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>>>>>> >>>>>>>>> For questions or feedback about the sbml-discuss list, >>>>>>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>>>>>> >>>> -- >>>> Herbert M. Sauro, PhD >>>> Associate Professor >>>> Department of Bioengineering >>>> >>>> University of Washington >>>> William H. Foege Building >>>> Box 355061, Seattle, WA 98195-5061 >>>> 206.685.2119 (tel), 206.685.3300 (shared fax) >>>> >>>> <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> >>>> New Text Book: Enzyme Kinetics for Systems Biology >>>> Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> >>>> >>>> ____________________________________________________________ >>>> To manage your sbml-discuss list subscription, visit >>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>> >>>> For a web interface to the sbml-discuss mailing list, visit >>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>> >>>> For questions or feedback about the sbml-discuss list, >>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>> ____________________________________________________________ >>> To manage your sbml-discuss list subscription, visit >>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>> >>> For a web interface to the sbml-discuss mailing list, visit >>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>> >>> For questions or feedback about the sbml-discuss list, >>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >> >> -- >> Herbert M. Sauro, PhD >> Associate Professor >> Department of Bioengineering >> >> University of Washington >> William H. Foege Building >> Box 355061, Seattle, WA 98195-5061 >> 206.685.2119 (tel), 206.685.3300 (shared fax) >> >> <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> >> New Text Book: Enzyme Kinetics for Systems Biology >> Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> >> >> ____________________________________________________________ >> To manage your sbml-discuss list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >> >> For a web interface to the sbml-discuss mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-discuss list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> -- Herbert M. Sauro, PhD Associate Professor Department of Bioengineering University of Washington William H. Foege Building Box 355061, Seattle, WA 98195-5061 206.685.2119 (tel), 206.685.3300 (shared fax) <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> New Text Book: Enzyme Kinetics for Systems Biology Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

Lucian Smith — 2012-09-06T18:28:03-00:00

Yes, the idea is that you would use libSBML. Different packages have different levels of libSBML support--for hierarchy, I tried to include a lot of example code so that getting started would be easier, and I believe a lot of the other packages have a good amount of example code as well. However, now that libSBML package support has matured enough, you're right that some sort of tutorial on how to code for packages in general would be great, along with more specific tutorials on coding for particular packages. Time is always the issue, but I bet getting something together for HARMONY would be possible, if there was enough interest. -Lucian * hsauro <<a href="mailto:hsauro@u.washington.edu" target="_blank">hsauro@u.washington.edu</a>> [2012-09-06 18:48] writes: > Its supporting the packages that scares me in terms of effort expended. > I tried looking up documentation on the sbml.org site this morning on > how to work with packages but I couldn't find anything other than a list > of packages in progress. I thought there was some docs somewhere. I > looked at the distrib package docs and found this little handy snip bit > from one of the examples: > > > > > > 3.14</un:logScale> > 3.14</un:shape> > </un:LogNormalDistribution> > </distrib:distribution> > > > mu</ci></bvar> > sigma</ci></bvar> > mu</ci> > </lambda> > </math> > </functionDefinition> > > What do I actually do with it? I want to avoid working through the XML > myself. Can I use libSBML to pull the information out without too much > effort? If its too hard to pull the info out, very few people are likely > use it other than those with the necessary human resources. > > Herbert > > > On 9/6/2012 10:10 AM, Chris J. Myers wrote: > > That is direction we are heading with distrib which I think should be ready very soon. Stuart and Sarah can comment on this. > > > > As for not moving to L3, this is not difficult as it is very similar to L2. You only need to deal with those packages you find useful. Switching to L3 for us was trivial. Making use of packages took some work, but it was really worth it as it allowed us to replace much of the specialized stuff we were doing in our tool that made our models non-exchangeable. Hierarchical modeling which is now complete and should have final approval this month should be incentive enough to move to L3, if you ask me. > > > > Chris > > > > On Sep 6, 2012, at 11:00 AM, hsauro wrote: > > > >> ok I see, thanks for the clarification. In that case one just needs to > >> include the uniform random number generator inside each user defined > >> distribution. There are however two downsides to this approach, the > >> random number generator can't be very complex (no looping in MathML, no > >> Mersenne twister?) and evaluation of the random number might be > >> inefficient at the tool end since it has to interpret the user defined > >> function. If we could have a built-in random number generator in SBML > >> just as we have sin, cos etc that would help a lot with this situation. > >> > >> Herbert Sauro > >> > >> On 9/6/2012 9:52 AM, Nicolas Le Nov?re wrote: > >>> Yes, one can do > >>> > >>> FunctionDef F1=ax+b > >>> FuntionDef F2=cy+d > >>> > >>> assignment F2(F1(z)) > >>> > >>> But one cannot do > >>> > >>> FunctionDef F1=ax+b > >>> FuntionDef F2=cF1(x)+d > >>> > >>> assignment F2(z) > >>> > >>> > >>> On 06/09/12 17:43, hsauro wrote: > >>>> I just checked the SBML spec for L2V4, couldn't figure out whether user > >>>> defined functions can call other user defined functions or not. It says > >>>> that cannot be recursive but I assume it is ok for one to call another > >>>> assuming no recursion? > >>>> > >>>> Herbert Sauro > >>>> > >>>> On 9/6/2012 9:35 AM, hsauro wrote: > >>>>> Nicolas has a point. However there is a simple way out of it. We just > >>>>> define our own uniform random number generator as a user defined > >>>>> function (we pick an algorithm we like), then declare a set of > >>>>> alternative distribution user defined functions that use the uniform > >>>>> random number generator to compute the given distribution. This saves > >>>>> work in having to upgrade to L3 and deal with packages which many of us > >>>>> don't have time to do at the moment. > >>>>> > >>>>> Herbert Sauro > >>>>> > >>>>> On 9/6/2012 9:15 AM, Nicolas Le Nov?re wrote: > >>>>>> Hi Frank, > >>>>>> > >>>>>> This is great that software start to exchange models with uncertainty. But > >>>>>> I am puzzled by your approach of developing your own annotation rather than > >>>>>> putting distrib constructs in the annotation element. That would be the > >>>>>> favoured way to provide an easier path towards the package once accepted, > >>>>>> and also minimise discrepancy between level 2 and level 3. That is at least > >>>>>> what has been done with the layout package. > >>>>>> > >>>>>> In effect what you propose is very close to distrib except you use > >>>>>> wikipedia instead of UncertML. However, since wikipedia cannot be used > >>>>>> easily programmaticaly, there is an uncertainty on the identity of the > >>>>>> parameter to feed the distribution (effectively the same problem we had > >>>>>> with the predefined functions of level 1). > >>>>>> > >>>>>> ? > >>>>>> > >>>>>> On 06/09/12 16:38, Frank T. Bergmann wrote: > >>>>>>> Hello everyone, > >>>>>>> > >>>>>>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). > >>>>>>> > >>>>>>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> </annotation> > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> m </ci> > >>>>>>> </bvar> > >>>>>>> > >>>>>>> s </ci> > >>>>>>> </bvar> > >>>>>>> m </ci> > >>>>>>> </lambda> > >>>>>>> </math> > >>>>>>> </functionDefinition> > >>>>>>> > >>>>>>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. > >>>>>>> > >>>>>>> more information and the specification here: > >>>>>>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> > >>>>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> > >>>>>>> > >>>>>>> Cheers > >>>>>>> Frank > >>>>>>> ____________________________________________________________ > >>>>>>> To manage your sbml-discuss list subscription, visit > >>>>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > >>>>>>> > >>>>>>> For a web interface to the sbml-discuss mailing list, visit > >>>>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > >>>>>>> > >>>>>>> For questions or feedback about the sbml-discuss list, > >>>>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > >>>>>>> > >> > >> -- > >> Herbert M. Sauro, PhD > >> Associate Professor > >> Department of Bioengineering > >> > >> University of Washington > >> William H. Foege Building > >> Box 355061, Seattle, WA 98195-5061 > >> 206.685.2119 (tel), 206.685.3300 (shared fax) > >> > >> <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> > >> New Text Book: Enzyme Kinetics for Systems Biology > >> Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> > >> > >> ____________________________________________________________ > >> To manage your sbml-discuss list subscription, visit > >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > >> > >> For a web interface to the sbml-discuss mailing list, visit > >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > >> > >> For questions or feedback about the sbml-discuss list, > >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > ____________________________________________________________ > > To manage your sbml-discuss list subscription, visit > > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > > > For a web interface to the sbml-discuss mailing list, visit > > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > > > For questions or feedback about the sbml-discuss list, > > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > > > -- > Herbert M. Sauro, PhD > Associate Professor > Department of Bioengineering > > University of Washington > William H. Foege Building > Box 355061, Seattle, WA 98195-5061 > 206.685.2119 (tel), 206.685.3300 (shared fax) > > <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> > New Text Book: Enzyme Kinetics for Systems Biology > Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> > > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

hsauro — 2012-09-06T17:36:18-00:00

Its supporting the packages that scares me in terms of effort expended. I tried looking up documentation on the sbml.org site this morning on how to work with packages but I couldn't find anything other than a list of packages in progress. I thought there was some docs somewhere. I looked at the distrib package docs and found this little handy snip bit from one of the examples: 3.14</un:logScale> 3.14</un:shape> </un:LogNormalDistribution> </distrib:distribution> mu</ci></bvar> sigma</ci></bvar> mu</ci> </lambda> </math> </functionDefinition> What do I actually do with it? I want to avoid working through the XML myself. Can I use libSBML to pull the information out without too much effort? If its too hard to pull the info out, very few people are likely use it other than those with the necessary human resources. Herbert On 9/6/2012 10:10 AM, Chris J. Myers wrote: > That is direction we are heading with distrib which I think should be ready very soon. Stuart and Sarah can comment on this. > > As for not moving to L3, this is not difficult as it is very similar to L2. You only need to deal with those packages you find useful. Switching to L3 for us was trivial. Making use of packages took some work, but it was really worth it as it allowed us to replace much of the specialized stuff we were doing in our tool that made our models non-exchangeable. Hierarchical modeling which is now complete and should have final approval this month should be incentive enough to move to L3, if you ask me. > > Chris > > On Sep 6, 2012, at 11:00 AM, hsauro wrote: > >> ok I see, thanks for the clarification. In that case one just needs to >> include the uniform random number generator inside each user defined >> distribution. There are however two downsides to this approach, the >> random number generator can't be very complex (no looping in MathML, no >> Mersenne twister?) and evaluation of the random number might be >> inefficient at the tool end since it has to interpret the user defined >> function. If we could have a built-in random number generator in SBML >> just as we have sin, cos etc that would help a lot with this situation. >> >> Herbert Sauro >> >> On 9/6/2012 9:52 AM, Nicolas Le Novère wrote: >>> Yes, one can do >>> >>> FunctionDef F1=ax+b >>> FuntionDef F2=cy+d >>> >>> assignment F2(F1(z)) >>> >>> But one cannot do >>> >>> FunctionDef F1=ax+b >>> FuntionDef F2=cF1(x)+d >>> >>> assignment F2(z) >>> >>> >>> On 06/09/12 17:43, hsauro wrote: >>>> I just checked the SBML spec for L2V4, couldn't figure out whether user >>>> defined functions can call other user defined functions or not. It says >>>> that cannot be recursive but I assume it is ok for one to call another >>>> assuming no recursion? >>>> >>>> Herbert Sauro >>>> >>>> On 9/6/2012 9:35 AM, hsauro wrote: >>>>> Nicolas has a point. However there is a simple way out of it. We just >>>>> define our own uniform random number generator as a user defined >>>>> function (we pick an algorithm we like), then declare a set of >>>>> alternative distribution user defined functions that use the uniform >>>>> random number generator to compute the given distribution. This saves >>>>> work in having to upgrade to L3 and deal with packages which many of us >>>>> don't have time to do at the moment. >>>>> >>>>> Herbert Sauro >>>>> >>>>> On 9/6/2012 9:15 AM, Nicolas Le Novère wrote: >>>>>> Hi Frank, >>>>>> >>>>>> This is great that software start to exchange models with uncertainty. But >>>>>> I am puzzled by your approach of developing your own annotation rather than >>>>>> putting distrib constructs in the annotation element. That would be the >>>>>> favoured way to provide an easier path towards the package once accepted, >>>>>> and also minimise discrepancy between level 2 and level 3. That is at least >>>>>> what has been done with the layout package. >>>>>> >>>>>> In effect what you propose is very close to distrib except you use >>>>>> wikipedia instead of UncertML. However, since wikipedia cannot be used >>>>>> easily programmaticaly, there is an uncertainty on the identity of the >>>>>> parameter to feed the distribution (effectively the same problem we had >>>>>> with the predefined functions of level 1). >>>>>> >>>>>> ? >>>>>> >>>>>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>>>>> Hello everyone, >>>>>>> >>>>>>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >>>>>>> >>>>>>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> </annotation> >>>>>>> >>>>>>> >>>>>>> >>>>>>> m </ci> >>>>>>> </bvar> >>>>>>> >>>>>>> s </ci> >>>>>>> </bvar> >>>>>>> m </ci> >>>>>>> </lambda> >>>>>>> </math> >>>>>>> </functionDefinition> >>>>>>> >>>>>>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >>>>>>> >>>>>>> more information and the specification here: >>>>>>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>>>>> >>>>>>> Cheers >>>>>>> Frank >>>>>>> ____________________________________________________________ >>>>>>> To manage your sbml-discuss list subscription, visit >>>>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>>>> >>>>>>> For a web interface to the sbml-discuss mailing list, visit >>>>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>>>> >>>>>>> For questions or feedback about the sbml-discuss list, >>>>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>>>> >> >> -- >> Herbert M. Sauro, PhD >> Associate Professor >> Department of Bioengineering >> >> University of Washington >> William H. Foege Building >> Box 355061, Seattle, WA 98195-5061 >> 206.685.2119 (tel), 206.685.3300 (shared fax) >> >> <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> >> New Text Book: Enzyme Kinetics for Systems Biology >> Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> >> >> ____________________________________________________________ >> To manage your sbml-discuss list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >> >> For a web interface to the sbml-discuss mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-discuss list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> -- Herbert M. Sauro, PhD Associate Professor Department of Bioengineering University of Washington William H. Foege Building Box 355061, Seattle, WA 98195-5061 206.685.2119 (tel), 206.685.3300 (shared fax) <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> New Text Book: Enzyme Kinetics for Systems Biology Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

myers — 2012-09-06T17:10:26-00:00

That is direction we are heading with distrib which I think should be ready very soon. Stuart and Sarah can comment on this. As for not moving to L3, this is not difficult as it is very similar to L2. You only need to deal with those packages you find useful. Switching to L3 for us was trivial. Making use of packages took some work, but it was really worth it as it allowed us to replace much of the specialized stuff we were doing in our tool that made our models non-exchangeable. Hierarchical modeling which is now complete and should have final approval this month should be incentive enough to move to L3, if you ask me. Chris On Sep 6, 2012, at 11:00 AM, hsauro wrote: > ok I see, thanks for the clarification. In that case one just needs to > include the uniform random number generator inside each user defined > distribution. There are however two downsides to this approach, the > random number generator can't be very complex (no looping in MathML, no > Mersenne twister?) and evaluation of the random number might be > inefficient at the tool end since it has to interpret the user defined > function. If we could have a built-in random number generator in SBML > just as we have sin, cos etc that would help a lot with this situation. > > Herbert Sauro > > On 9/6/2012 9:52 AM, Nicolas Le Novère wrote: >> Yes, one can do >> >> FunctionDef F1=ax+b >> FuntionDef F2=cy+d >> >> assignment F2(F1(z)) >> >> But one cannot do >> >> FunctionDef F1=ax+b >> FuntionDef F2=cF1(x)+d >> >> assignment F2(z) >> >> >> On 06/09/12 17:43, hsauro wrote: >>> I just checked the SBML spec for L2V4, couldn't figure out whether user >>> defined functions can call other user defined functions or not. It says >>> that cannot be recursive but I assume it is ok for one to call another >>> assuming no recursion? >>> >>> Herbert Sauro >>> >>> On 9/6/2012 9:35 AM, hsauro wrote: >>>> Nicolas has a point. However there is a simple way out of it. We just >>>> define our own uniform random number generator as a user defined >>>> function (we pick an algorithm we like), then declare a set of >>>> alternative distribution user defined functions that use the uniform >>>> random number generator to compute the given distribution. This saves >>>> work in having to upgrade to L3 and deal with packages which many of us >>>> don't have time to do at the moment. >>>> >>>> Herbert Sauro >>>> >>>> On 9/6/2012 9:15 AM, Nicolas Le Novère wrote: >>>>> Hi Frank, >>>>> >>>>> This is great that software start to exchange models with uncertainty. But >>>>> I am puzzled by your approach of developing your own annotation rather than >>>>> putting distrib constructs in the annotation element. That would be the >>>>> favoured way to provide an easier path towards the package once accepted, >>>>> and also minimise discrepancy between level 2 and level 3. That is at least >>>>> what has been done with the layout package. >>>>> >>>>> In effect what you propose is very close to distrib except you use >>>>> wikipedia instead of UncertML. However, since wikipedia cannot be used >>>>> easily programmaticaly, there is an uncertainty on the identity of the >>>>> parameter to feed the distribution (effectively the same problem we had >>>>> with the predefined functions of level 1). >>>>> >>>>> ? >>>>> >>>>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>>>> Hello everyone, >>>>>> >>>>>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >>>>>> >>>>>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> </annotation> >>>>>> >>>>>> >>>>>> >>>>>> m </ci> >>>>>> </bvar> >>>>>> >>>>>> s </ci> >>>>>> </bvar> >>>>>> m </ci> >>>>>> </lambda> >>>>>> </math> >>>>>> </functionDefinition> >>>>>> >>>>>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >>>>>> >>>>>> more information and the specification here: >>>>>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>>>> >>>>>> Cheers >>>>>> Frank >>>>>> ____________________________________________________________ >>>>>> To manage your sbml-discuss list subscription, visit >>>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>>> >>>>>> For a web interface to the sbml-discuss mailing list, visit >>>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>>> >>>>>> For questions or feedback about the sbml-discuss list, >>>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>>> >>> >> > > > -- > Herbert M. Sauro, PhD > Associate Professor > Department of Bioengineering > > University of Washington > William H. Foege Building > Box 355061, Seattle, WA 98195-5061 > 206.685.2119 (tel), 206.685.3300 (shared fax) > > <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> > New Text Book: Enzyme Kinetics for Systems Biology > Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> > > ____________________________________________________________ > To manage your sbml-discuss list subscription, visit > <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> > > For a web interface to the sbml-discuss mailing list, visit > <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> > > For questions or feedback about the sbml-discuss list, > contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

hsauro — 2012-09-06T17:09:18-00:00

Ah yes, I see that now. So it is possible. Herbert On 9/6/2012 9:59 AM, Lucian Smith wrote: > That is, the contents of MathML ci elements inside the body of the > lambda can only be the variables declared by its bvar elements, or the > identifiers of other FunctionDefinitions defined in the same model -- Herbert M. Sauro, PhD Associate Professor Department of Bioengineering University of Washington William H. Foege Building Box 355061, Seattle, WA 98195-5061 206.685.2119 (tel), 206.685.3300 (shared fax) <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> New Text Book: Enzyme Kinetics for Systems Biology Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: Custom Annotation for drawing numbers from Distributions

hsauro — 2012-09-06T17:00:56-00:00

ok I see, thanks for the clarification. In that case one just needs to include the uniform random number generator inside each user defined distribution. There are however two downsides to this approach, the random number generator can't be very complex (no looping in MathML, no Mersenne twister?) and evaluation of the random number might be inefficient at the tool end since it has to interpret the user defined function. If we could have a built-in random number generator in SBML just as we have sin, cos etc that would help a lot with this situation. Herbert Sauro On 9/6/2012 9:52 AM, Nicolas Le Novère wrote: > Yes, one can do > > FunctionDef F1=ax+b > FuntionDef F2=cy+d > > assignment F2(F1(z)) > > But one cannot do > > FunctionDef F1=ax+b > FuntionDef F2=cF1(x)+d > > assignment F2(z) > > > On 06/09/12 17:43, hsauro wrote: >> I just checked the SBML spec for L2V4, couldn't figure out whether user >> defined functions can call other user defined functions or not. It says >> that cannot be recursive but I assume it is ok for one to call another >> assuming no recursion? >> >> Herbert Sauro >> >> On 9/6/2012 9:35 AM, hsauro wrote: >>> Nicolas has a point. However there is a simple way out of it. We just >>> define our own uniform random number generator as a user defined >>> function (we pick an algorithm we like), then declare a set of >>> alternative distribution user defined functions that use the uniform >>> random number generator to compute the given distribution. This saves >>> work in having to upgrade to L3 and deal with packages which many of us >>> don't have time to do at the moment. >>> >>> Herbert Sauro >>> >>> On 9/6/2012 9:15 AM, Nicolas Le Novère wrote: >>>> Hi Frank, >>>> >>>> This is great that software start to exchange models with uncertainty. But >>>> I am puzzled by your approach of developing your own annotation rather than >>>> putting distrib constructs in the annotation element. That would be the >>>> favoured way to provide an easier path towards the package once accepted, >>>> and also minimise discrepancy between level 2 and level 3. That is at least >>>> what has been done with the layout package. >>>> >>>> In effect what you propose is very close to distrib except you use >>>> wikipedia instead of UncertML. However, since wikipedia cannot be used >>>> easily programmaticaly, there is an uncertainty on the identity of the >>>> parameter to feed the distribution (effectively the same problem we had >>>> with the predefined functions of level 1). >>>> >>>> ? >>>> >>>> On 06/09/12 16:38, Frank T. Bergmann wrote: >>>>> Hello everyone, >>>>> >>>>> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >>>>> >>>>> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >>>>> >>>>> >>>>> >>>>> >>>>> </annotation> >>>>> >>>>> >>>>> >>>>> m </ci> >>>>> </bvar> >>>>> >>>>> s </ci> >>>>> </bvar> >>>>> m </ci> >>>>> </lambda> >>>>> </math> >>>>> </functionDefinition> >>>>> >>>>> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >>>>> >>>>> more information and the specification here: >>>>> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >>>>> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >>>>> >>>>> Cheers >>>>> Frank >>>>> ____________________________________________________________ >>>>> To manage your sbml-discuss list subscription, visit >>>>> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >>>>> >>>>> For a web interface to the sbml-discuss mailing list, visit >>>>> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >>>>> >>>>> For questions or feedback about the sbml-discuss list, >>>>> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >>>>> >> > -- Herbert M. Sauro, PhD Associate Professor Department of Bioengineering University of Washington William H. Foege Building Box 355061, Seattle, WA 98195-5061 206.685.2119 (tel), 206.685.3300 (shared fax) <a href="http://www.sys-bio.org" target="_blank">http://www.sys-bio.org</a>; <a href="http://blog.analogmachine.org" target="_blank">http://blog.analogmachine.org</a> New Text Book: Enzyme Kinetics for Systems Biology Order at Amazon: <a href="http://tinyurl.com/3w4b9al" target="_blank">http://tinyurl.com/3w4b9al</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>

Re: [sbml-distrib] Custom Annotation for drawing numbers from Distributions

myers — 2012-09-06T17:00:11-00:00

I agree with Nicolas. I think distrib is very close, and you can make an annotation scheme for L2 which is similar to what we have agreed on with distrib. Stuart: do you have a specification which is up-to-date with respect to our decisions in Toronto that you can send to Frank? Chris On Sep 6, 2012, at 10:15 AM, Nicolas Le Novère wrote: > Hi Frank, > > This is great that software start to exchange models with uncertainty. But > I am puzzled by your approach of developing your own annotation rather than > putting distrib constructs in the annotation element. That would be the > favoured way to provide an easier path towards the package once accepted, > and also minimise discrepancy between level 2 and level 3. That is at least > what has been done with the layout package. > > In effect what you propose is very close to distrib except you use > wikipedia instead of UncertML. However, since wikipedia cannot be used > easily programmaticaly, there is an uncertainty on the identity of the > parameter to feed the distribution (effectively the same problem we had > with the predefined functions of level 1). > > ? > > On 06/09/12 16:38, Frank T. Bergmann wrote: >> Hello everyone, >> >> I'm writing today to make you aware of a custom annotation for SBML Level 2 (or Level 3) models, that makes it possible to draw random numbers from a number of distributions. This is a feature many of us have encountered several times, but that is not yet standardized (until the release of the L3 distrib package). >> >> This is done by simply tagging a function definition with the right number of arguments, to let us know that this indeed is a distribution function. Here just one example: >> >> >> >> >> </annotation> >> >> >> >> m </ci> >> </bvar> >> >> s </ci> >> </bvar> >> m </ci> >> </lambda> >> </math> >> </functionDefinition> >> >> By identifying the distribution function via the definition tag, it no longer matters how a tool names the function, making it easier to share across tools. Please let me know if you have any questions or concerns, or if you would like to implement support for it in your tool and require more information. >> >> more information and the specification here: >> - <a href="http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html" target="_blank">http://frank-fbergmann.blogspot.com/2012/09/custom-annotation-for-using.html</a> >> - <a href="https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1" target="_blank">https://docs.google.com/file/d/0B_wMqVOQLkZ3TVZHblNNRWgzNTg/edit?pli=1</a> >> >> Cheers >> Frank >> ____________________________________________________________ >> To manage your sbml-discuss list subscription, visit >> <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> >> >> For a web interface to the sbml-discuss mailing list, visit >> <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> >> >> For questions or feedback about the sbml-discuss list, >> contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a> >> > > > -- > Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC, > Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468, > <a href="mailto:lenov@ebi.ac.uk" target="_blank">lenov@ebi.ac.uk</a>, Skype:n.lenovere, twitter:@lenovere > <a href="http://www.ebi.ac.uk/~lenov/," target="_blank">http://www.ebi.ac.uk/~lenov/,</a> <a href="http://lenoverelab.org/" target="_blank">http://lenoverelab.org/</a> > > > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. <a href="http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/" target="_blank">http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/</a> > _______________________________________________ > sbml-distrib mailing list > <a href="mailto:sbml-distrib@lists.sourceforge.net" target="_blank">sbml-distrib@lists.sourceforge.net</a> > <a href="https://lists.sourceforge.net/lists/listinfo/sbml-distrib" target="_blank">https://lists.sourceforge.net/lists/listinfo/sbml-distrib</a> ____________________________________________________________ To manage your sbml-discuss list subscription, visit <a href="https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss" target="_blank">https://utils.its.caltech.edu/mailman/listinfo/sbml-discuss</a> For a web interface to the sbml-discuss mailing list, visit <a href="http://sbml.org/Forums/" target="_blank">http://sbml.org/Forums/</a> For questions or feedback about the sbml-discuss list, contact <a href="mailto:sbml-team@caltech.edu" target="_blank">sbml-team@caltech.edu</a>