Forums

F.A.Q. F.A.Q.    Register Register    Login Login    Home Home
Search Search
SBML Discussions » jsbml-development » Bug in SBMLWriter
Show: Today's Posts  :: Message Navigator
| Subscribe to topic 
Return to the default flat view Create a new topic Submit Reply
AuthorTopic
rodrigue


Posts: 307
Location:
Cambridge UK
Registered:
February 2005
Re: Bug in SBMLWriter 10 May '12 05:38 Go to previous messageGo to previous message

On 10/05/12 12:57, Nicolas Rodriguez wrote:
> On 10/05/12 12:49, Nicolas Rodriguez wrote:
>> On 09/05/12 19:42, Andreas Dräger wrote:
>>> Am 5/8/12 3:43 PM, schrieb Nicolas Rodriguez:
>>>> On 08/05/12 14:30, Nicolas Rodriguez wrote:
>>>>> On 08/05/12 14:20, Andreas Dräger wrote:
>>>>>> Am 5/8/12 3:17 PM, schrieb Nicolas Rodriguez:
>>>>>>> On 08/05/12 13:57, Andreas Dräger wrote:
>>>>>>>> |Hi all,
>>>>>>>>
>>>>>>>> It seems the SBML writer does currently write invalid SBML code when
>>>>>>>> making use of the ability of SBML Level 3 to associate plain numbers
>>>>>>>> with units:
>>>>>>>>
>>>>>>>> <cn type="integer" sbml:units="dimensionless"> 1</cn>
>>>>>>>>
>>>>>>>> The above code is only valid if the namespace declaration of the math
>>>>>>>> element also defines
>>>>>>>>
>>>>>>>> xmlns:sbml="http://www.sbml.org/sbml/level3/version1/core"
>>>>>>>>
>>>>>>>> Currently, it does only declare the regular MathML namesapce. As a
>>>>>>>> result, JSBML is currently unable to parse models that itself wrote into
>>>>>>>> a file before. Where can we fix this problem?
>>>>>>> The check is in
>>>>>>> org.sbml.jsbml.util.compilers.MathMLXMLStreamCompiler.isSBMLNamespaceNeeded(ASTNode)
>>>>>>>
>>>>>>> the namespace is added in
>>>>>>> org.sbml.jsbml.xml.stax.SBMLWriter.writeMathML(MathContainer,
>>>>>>> SMOutputElement, XMLStreamWriter, int)
>>>>>>>
>>>>>>> line 1084 to 1106.
>>>>>>>
>>>>>>> The compiler org.sbml.jsbml.util.compilers.FindUnitsCompiler is used to
>>>>>>> say if there is a unit declaration or not in the mathML.
>>>>>>>
>>>>>>> This is probably there that the problem happen, and the value returned
>>>>>>> is probably false.
>>>>>>> It is probably due to the way you are declaring the units and then how
>>>>>>> the mathML compiler is calling the methods.
>>>>>>>
>>>>>>> If I read a model with units, then the writing is fine so there is may
>>>>>>> be something missing when we declare units from scratch.
>>>>>>>
>>>>>>>
>>>>>> Hi Nico,
>>>>>>
>>>>>> Yes, it is probably because I simply say
>>>>>>
>>>>>> myASTNode.setUnit(Unit.Kind.DIMENSIONLESS.getName());
>>>>>>
>>>>>> or something similar. I thought this would be enough and didn't care
>>>>>> about namespaces at the moment.
>>>>> It should be enough but it was probably not tested much until now.
>>>>>
>>>>> We should probably check how the reader is setting things up.
>>>> So, the reader is not doing anything specific, just using the setUnits()
>>>> method ( I suppose it was a typo error on your code as I don't see a
>>>> method setUnit().
>>>>
>>>> Strange, it was working fine before.
>>>>
>>>>
>>> Hi all,
>>>
>>> here is a minimal example that doesn't work properly with the current
>>> version of JSBML:
>>>
>>> public static void main(String[] args) throws Throwable {
>>> SBMLDocument doc = new SBMLDocument(3, 1);
>>> Model m = doc.createModel("m1");
>>> Compartment c = m.createCompartment("c1");
>>> Species s1 = m.createSpecies("s1", c);
>>> Species s2 = m.createSpecies("s2", c);
>>> Reaction r = m.createReaction("r1");
>>> r.createReactant(s1).setStoichiometry(1d);
>>> r.createProduct(s2).setStoichiometry(1d);
>>> KineticLaw kl = r.createKineticLaw();
>>> ASTNode one = new ASTNode(1, kl);
>>> ASTNode two = new ASTNode(2, kl);
>>> one.setUnits(Unit.Kind.MOLE.toString().toLowerCase());
>>> two.setUnits(Unit.Kind.SECOND.toString().toLowerCase());
>>> ASTNode ast = ASTNode.frac(one, two);
>>> kl.setMath(ast);
>>> SBMLWriter.write(doc, System.out, ' ', (short) 2);
>>> }
>>>
>>> You'll see that the first unit is not written to SBML at all. The second
>>> unit is written, but the namespace declaration is missing. I already
>>> found what is probably the reason, but can't fix it at the moment: The
>>> recursive check if the ASTNode contains unit declarations doesn't work.
>> Not sure of what is going on there !!
>>
>> The unit of the first astnode is reset somewhere.
>>
>> if I print the value of the units in
>> MathMLXMLStreamCompiler.compileInteger(ASTNode), the first astnode
>> return 'null'.
> The unit that get deleted is probably due to the code from
> org.sbml.jsbml.ASTNode.arithmeticOperation(Type, ASTNode) ??
>
> The swap node does not copy the units ?

ok, both fixes are commited now and it should be working fine.

I suppose this a good case to de-freeze the 0.8 branch.

Should I commit there (in the 0.8 branch) the changes to
Model.registerIds as well ?

Nico


____________________________________________________________
To manage your jsbml-development list subscription, visit
https://utils.its.caltech.edu/mailman/listinfo/jsbml-development

For a web interface to the jsbml-development mailing list, visit
http://sbml.org/Forums/

For questions or feedback about the jsbml-development list,
contact sbml-team@caltech.edu

      

SubjectPosterDate
Read Message   Bug in SBMLWriter Andreas Dräger08 May '12 05:57
Read Message   Re: Bug in SBMLWriter rodrigue08 May '12 06:17
Read Message   Re: Bug in SBMLWriter Andreas Dräger08 May '12 06:20
Read Message   Re: Bug in SBMLWriter rodrigue08 May '12 06:30
Read Message   Re: Bug in SBMLWriter rodrigue08 May '12 06:43
Read Message   Re: Bug in SBMLWriter rodrigue08 May '12 07:10
Read Message   Re: Bug in SBMLWriter Andreas Dräger09 May '12 11:42
Read Message   Re: Bug in SBMLWriter rodrigue10 May '12 04:49
Read Message   Re: Bug in SBMLWriter rodrigue10 May '12 04:57
Read Message   Re: Bug in SBMLWriter  rodrigue10 May '12 05:38
Read Message   Re: Bug in SBMLWriter Andreas Dräger10 May '12 05:50
Read Message   Re: Bug in SBMLWriter Andreas Dräger10 May '12 04:59
Previous Topic:Re: [sbml-render] Ids and names on elements
Next Topic:jsbml-svn commit list added
Go to forum:
-=] Back to Top [=-

Powered by FUDforum. (Copyright Advanced Internet Designs Inc.)

Please use our issue tracking system for any questions or suggestions about this website.