From: Dagmar K. <da...@eb...> - 2009-03-22 11:45:28
|
Just to say that I agree with everything Frank wrote (hopefully I understood right what he meant ;-)). Dagmar Frank Bergmann wrote: > Hello ... > > >> Hi, >> Just a query about the listOfModels element - is it assumed to be >> the case that, for a given sedml document, it will describe a single >> 'reference' model version, which will be the first model in the list, >> and subsequent entries in the list will refer to that base model, >> applying changes (as is the case with the example 'oscillation to >> chaos' model)? Or, is it possible that a given sedml document could >> refer to >1 reference model? >> > > Definitely ... you can (and should be able to) refer to more than one model. > > > >> I'm asking from the point of view of how a self-contained .miase >> archive file might look. (By self contained, I mean that the >> application using it will be able to access all the information needed >> to run a simulation without having to query repositories - with the >> help of a supporting library. ) In the first case above, it would >> only ever contain a single model file, in the second case, it could >> contain multiple models. >> > > In my implementation, the self-contained .miase file, still has the option > of referring to external sources. Thus There could be 0 or more models > contained in the archive. > > >> Also, do we want to allow archives which contain all model variants as >> complete models, or are we going to stick with a 'reference' version >> of the model, and generate new, altered versions of the model by >> applying Xpath changes on the fly at runtime? I assume the latter, but >> I can't find anything in the current examples, or in the CMSB2008.pdf >> to distinguish between these possibilities. >> >> > > I hoped to keep the .miase file as simple as possible. It really just > provides the option of storing (locally) referenced files conveniently. > Apart from that there should not be more logic needed to interpret a .miase > file, as compared to the original experiment description. For that reason > I'd opt of making the changes on the fly, they won't be expensive to do. > While writing out you can opt to refer to the altered models if you so > desire, however with that you loose the information, that the second model > is actually the first model with a handful of changes ... > > >> From the point of view of supporting library methods, I was thinking >> something along the lines of: >> >> MiaseArchive archive = libSedml.readArchive(String >> filepathToMiaseArchive); >> >> // A MiaseArchive enables access to a sedml document, and model >> file(s?) >> SedmlDocument document = archive.getSedMlDocument(); >> // access Sedml DOM via >> document.getSedMlModel(); >> >> //access model file ( or is it possible to have > 1 model file >> here??) >> > > YES it is ... :) ... > > >> // if the latter, need method to iterate over multiple model files. >> String referenceModelContents = archive.getModelAsString(); >> >> //now, get XPath change from SedML model and apply it >> ListOfChanges changes = >> sedMl.getListOfModels.getModelById(modelID).getListOfChanges(); >> String alteredModelContents = >> document.applyChanges(referenceModelContents,changes); >> >> > > Instead of > > document.applyChanges(referenceModelContents,changes); > > I'd prefer to just grab the changed model. That's what I did in my > implementation, for a user there is hardly a difference on the two model > types. Thy have the option of getting the model source from both, and in > that case the changed model, applies the changes automatically. > > Best > Frank > > > ------------------------------------------------------------------------------ > Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are > powering Web 2.0 with engaging, cross-platform capabilities. Quickly and > easily build your RIAs with Flex Builder, the Eclipse(TM)based development > software that enables intelligent coding and step-through debugging. > Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com > _______________________________________________ > Miase-discuss mailing list > Mia...@li... > https://lists.sourceforge.net/lists/listinfo/miase-discuss > |