Menu
▾
▴
sbml2latex-development — Developer's mailing list
You can subscribe to this list here.
| 2013 |
Jan
|
Feb
|
Mar
(1) |
Apr
(5) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|---|
|
From: Nicolas R. <rod...@eb...> - 2013-04-11 15:49:46
|
On 04/11/2013 04:37 PM, Gabriel Kind wrote: > Hello, > > I researched this further by searching through the SBML L2V4 > specification (the sbml files I have to work with are based on L2). > The correct order is <math> followed by <listOfParameters> (different > order is invalid) and both are mandatory. > I tried to resolve this by adding an empty <math/> element but then the > MathML validation fails because empty formulas are invalid. Really great... > > Then I encountered that post on the SBML mailing list: > http://sbml.org/Forums/index.php?t=tree&th=1747 > > They suggest stuff like a fake formula that returns NaN as a result... > So I will stick to ommiting the math-element (or use an empty one) and > assume that the software I use is tolerant enough to handle this. :) You could have a look at what the COBRA toolbox is generating for example. Putting a math element with just 0 should do the trick for the validation. <math xmlns="http://www.w3.org/1998/Math/MathML"> <cn type="integer"> 0 </cn> </math> It all depends if you intend to exchange you model or not. If not, you could keep it without math, if yes, it would be better to define a kind of empty math. Nico |
|
From: Gabriel K. <gk...@hs...> - 2013-04-11 15:36:59
|
Hello, I researched this further by searching through the SBML L2V4 specification (the sbml files I have to work with are based on L2). The correct order is <math> followed by <listOfParameters> (different order is invalid) and both are mandatory. I tried to resolve this by adding an empty <math/> element but then the MathML validation fails because empty formulas are invalid. Really great... Then I encountered that post on the SBML mailing list: http://sbml.org/Forums/index.php?t=tree&th=1747 They suggest stuff like a fake formula that returns NaN as a result... So I will stick to ommiting the math-element (or use an empty one) and assume that the software I use is tolerant enough to handle this. :) ciao Gabriel |
|
From: Nicolas R. <rod...@eb...> - 2013-04-11 14:18:57
|
On 04/11/2013 03:14 PM, Andreas Dräger wrote:
> Am 4/11/13 4:02 PM, schrieb Gabriel Kind:
>> Can be resolved by changing
>> LatexReportGenerator.java:4002 from
>> if ((reaction == null) || !reaction.isSetKineticLaw()) {
>> to
>> if ((reaction == null) || !reaction.isSetKineticLaw() ||
>> !reaction.getKineticLaw().isSetMath()) {
>>
> Dear Gabriel,
>
> Thanks for reporting that! I just fixed the problem. Sometimes it is
> really hard to catch such errors.
>
>> After changing this the program reports that TeX compilation failed
>> but a PDF-Document is correctly generated (so that error message is
>> wrong? tex file attached)
>>
>> In the PDF it reports:
>> (10103 [Error]) [...] A KineticLaw object must contain exactly one
>> MathML <math> element.
>>
>> I validated using libsbml for Python and it reported 0 errors but
>> after reading the documentation of libsbml again I noticed that I have
>> to call document.checkConsistency() to get more errors. My bad.
> Used here is the online validator of libSBML via JSBML. This uses the
> latest version of libSBML. I am not sure which version you used, but I
> think, if you define a list of local parameters, you also have to have a
> math element. Actually, It is not clear to me why this is mandatory.
> Maybe, this question could be directed to some other mailing list, such
> as SBML discuss?
I think this was discussed quite a lot already, I am not sure what is
required in L3V1.
Nico
|
|
From: Andreas D. <and...@un...> - 2013-04-11 14:14:22
|
Am 4/11/13 4:02 PM, schrieb Gabriel Kind:
> Can be resolved by changing
> LatexReportGenerator.java:4002 from
> if ((reaction == null) || !reaction.isSetKineticLaw()) {
> to
> if ((reaction == null) || !reaction.isSetKineticLaw() ||
> !reaction.getKineticLaw().isSetMath()) {
>
Dear Gabriel,
Thanks for reporting that! I just fixed the problem. Sometimes it is
really hard to catch such errors.
> After changing this the program reports that TeX compilation failed
> but a PDF-Document is correctly generated (so that error message is
> wrong? tex file attached)
>
> In the PDF it reports:
> (10103 [Error]) [...] A KineticLaw object must contain exactly one
> MathML <math> element.
>
> I validated using libsbml for Python and it reported 0 errors but
> after reading the documentation of libsbml again I noticed that I have
> to call document.checkConsistency() to get more errors. My bad.
Used here is the online validator of libSBML via JSBML. This uses the
latest version of libSBML. I am not sure which version you used, but I
think, if you define a list of local parameters, you also have to have a
math element. Actually, It is not clear to me why this is mandatory.
Maybe, this question could be directed to some other mailing list, such
as SBML discuss?
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
|
|
From: Gabriel K. <gk...@hs...> - 2013-04-11 14:01:30
|
Dear developers,
I encountered a NullPointerException during conversion of one of my
SBML files.
The exception occurs at
org.sbml.tolatex.io.LaTeXReportGenerator.formatVelocity(LaTeXReportGenerator.java:4005)
(v.append(reaction.getKineticLaw().getMath().toLaTeX());)
getMath() returns a Null pointer because the <kinecticLaw> section in my
file has no <math> section. (attached a minimal XML sample file)
Can be resolved by changing
LatexReportGenerator.java:4002 from
if ((reaction == null) || !reaction.isSetKineticLaw()) {
to
if ((reaction == null) || !reaction.isSetKineticLaw() ||
!reaction.getKineticLaw().isSetMath()) {
After changing this the program reports that TeX compilation failed but
a PDF-Document is correctly generated (so that error message is wrong?
tex file attached)
In the PDF it reports:
(10103 [Error]) [...] A KineticLaw object must contain exactly one
MathML <math> element.
I validated using libsbml for Python and it reported 0 errors but after
reading the documentation of libsbml again I noticed that I have to call
document.checkConsistency() to get more errors. My bad.
So the fault is actually my SBML file and not the converter but maybe
this small fix is fine to be more tolerant against malformed SBML files
if it has no unwanted side effects.
ciao
Gabriel
|
|
From: Andreas D. <and...@un...> - 2013-03-25 07:47:16
|
Dear all, As wished by Camille and Nico, we have successfully moved the SBML2LaTeX project to sourceforge.net as its new host. You all should have access to https://sourceforge.net/projects/sbml2latex/. Those of you who would still like to participate in the development of this program are invited to also register for the mailing list (just click "Mailing Lists" at the top of the sourceforge page). For now, I included all team members to this mail under CC. Today, I comited a revision of the project, including the TeX files that were used to create the logo images of SBML2LaTeX and set all the SVN keywords. Furthermore, I updated and renamed the library SysBio*.jar to zbit.jar for the sake of simplicity and to indicate where it comes from. Its manifest file still contains further details. With the version number of the JAR file not being included in its name (but in its manifest file) it is easier to exchange this jar file more often without breaking the build path configuration of all team members. Recently, I was working on restructuring the project. The original version of SBML2LaTeX was basically one large Java class that contained LaTeX code and all the human-readable text to be included in the TeX document to be created based on the current SBML file. However, that's not very flexible and hard to maintain. I therefore decided to separate the creation of the report from the text and also from the LaTeX commands. I wanted to have all the text being included in easily exchangeable XML files and all the formatting commands in a formatter class. This would allow us to * implement formatters for different output formats, e.g., plain text or HTML * add translations of the text for several natural languages * spell-check and update the text more easily. With the implementation of SBML2TikZ in SBML2LaTeX, support for the first SBML extension package, layout, has been introduced as well. The integration of the layout into the model reports is not very good at the moment. On the long term, further SBML extensions should be included as well. Hence, a better concept is required to deal with extensions. I think, each such extension might declare several definitions in the document's header, maybe add a statement to the introduction section of the document, and finally add its actual content somewhere to the report. This means, there should be a more general interface or abstract class that could be used to ask the current extension package if * there is some report generator available for this package * it has something to declare in the report's header * it wants to add a statement to the introduction * it has some content to add to the report. Furthermore, an option should be added per supported extension class to decide whether the user wants to include a section about it in the report. For layout, such an option is already present in the interface LaTeXOptions. Please note that these options are used to automatically create the graphical user interface of SBML2LaTeX when launching the main class SBML2LaTeX. Please let me know if you have further wishes, concerns, ideas etc. 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 |
