[Miase-discuss] sed-ml library?

[Richard A. <ra...@st...>]

>
> ------------------------------
>
> Message: 6
> Date: Fri, 23 Jan 2009 17:35:11 +0100
> From: Dagmar K?hn <da...@eb...>
> Subject: Re: [Miase-discuss] sed-ml library?
> To: mia...@li...
> Message-ID: <497...@eb...>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
> Hej,
> regarding some of the issues:
>
>>    A Schema issues:
>>    --------------
>>       1.  According to the spec SEDBase is a superclass of all the main
>> classes, yest in the schema only SedML class inherits from it. This
>> means that notes and annotation cannot be added to other elements,
>> e.g., uniform time course, dataset, datagenerator etc.,
>>
>
> In the XML Schema it is missing indeed. I will add them, probably the
> easiest way is by adding the <xs:extension base="SEDBase"> tag to (all)
> complex type definitions.
>
>>       2. Most if not all top-level elements have an 'id' attribute,
>> would it make sense to put the 'id' field in SEDBase?
>>
>
> Well, what for example about the <curve> element which might have an
> annotation "The following curve shows the osciallation of bla bla bla",
> but does not have an id!?

  Do you mean that a curve should have an ID or shouldn't? If an id is  
going to be problematic to fit into an inheritsnce hierarcy, then  
probably it should juet be included where needed, but if there is an  
argument for all elements having an ID then it would make sense to  
have id in a common base class, like in sbml.

  One solution would be to add another base class with an 'id'  
attribute which extends SedBase ,so you have

                          SedBase
                        /         \
       IdentifiableSedBase         \
          /             \            \
       Simulation,   Output           Curve


  Or, if a complex inheritance hierarchy is preferred to be avoided,  
then an 'Annotation' class  could be defined which is composed into  
the top level elements, and a base class just has an ID.

>
>>    B)    Coordination with model
>>    ----------
>>      PRevious posts have raised this issue, is there any progress on
>> this? There is a lot of validation that can be done just with the
>> sedml file, but if we want to check that a parameter or variable name
>> exists, or the xpath is correct,
>>    then we need access to the model. It seems to me that there are
>> different scenarios:
>>

>
>>
>>          2) A model is not yet in the public domain, but for example
>> the client wants to share  simulation info with a collaborator. In
>> this case probably the model and  the sedml file would have to sent
>> together in a zip file to preserve
>>    relative file paths, but this leaves quite an onus on the user or
>> tool to keep the two files coordinated.
>>
>
> Hmm. This is indeed not really satisfying. Probably something to  think
> about.
>
>>          3) Could a model file be optionally  embedded in a sedml file?
>>
>
> For the moment, no. And  I would really like to avoid that. But if
> people want to do that they could wrap some additional XML element
> around a model and a sed-ml file... :-)
>
>>    This gets round the need to keep sedml independent of sbml, also a
>> single file gets round the file coordination issue,  but it raises new
>> problems of keeping in synch possibly multiple copies of a model. But
>> for non-curated models that must be an issue anyway?
>>
  I'll try and expland. I was thinking of a case of a non-curated  
model. For a ccurated model it is in a sense fixed, can be referred  
to, changes can be relative to it because it, or at least that version  
of the model is immutable. MAybe I don't understand the model creation  
process very well, but if you have different sedml files with embedded  
sbml files, each referring to a  different version of a particular  
model, then it is confusing to the user to have lots of slightly  
different model files in their system. But maybe htis isn't actually a  
problem, as it is up to a tool developer to provide support to the  
user, and I suppose from a library developlment point of view, so long  
as the combined sedml/sbml file is internally consistent, then that's  
the important thing.


>> C Validation requirements
>> --------------------------
>>
>>    VAlidation against the schema offers one level of validation, but
>> there are many more assertions to validate, here are
>> some below, would be good to have a list of these somewhere to
>> construct a test-suite.
>> Validation tasks below with an asterisk are ones needing access to   
>> the model:
>>
>
> As you said in an earlier email, I think the syntax checker is
> sufficient for the moment and the following validation criteria might be
> for the future!? However, this list could go to the feature request on SF.

OK I'll transfer it across. Some basic checkers( eg for uniform time  
course st and end time checking) are implemented just now, but need to  
work out how get the line numbers back post parsing.


Re Frank's comments about a C/C++ library and ported versions being  
the ideal, I can't disagree with that based on the success of libsbml,  
but  I do think there is room for a native Java implementation as well  
-   the Java bioinformatics community is quite large and just looking  
at the sbml interopperability forum there are a lot of questions about  
linking in the Java bindings. Just being able to drop a jar into your  
classpath would be quite handy for a lot of Java apps especially as  
often incorporating libsbml causes portability problems for what were  
previously  completley Xplatform apps.

I think what I'll try to do next is set up some more automated testing  
system so that as new example sedml afiles are added to svn they can  
be validated easily. Also need to think about design of aPI, code  
generation is great for getting started but the api is rather ungainly  
at present.


Just one more issue, how big do you think these files are likely to  
get, realistically, over the next few years? I'm thinking that as most  
models are relatively small then the sed-ml files will be easily held  
in memory for the forseeable future?

Cheers

Richard



-- 
Dr Richard Adams
Senior Software Developer,
Computational Systems Biology Group,
University of Edinburgh
Tel: 0131 650 8281/8285
email : ric...@ed...

-- 
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.