- From: david.nickerson at gmail.com (David Nickerson)
- Subject: [cellml-dev] Errors in CellML files
- Date: Fri, 23 Sep 2011 09:55:56 +1200
Hi Lucian,
Just to be clear, are you saying that your code hasn't changed but
something in the new 1.10 release of the CellML API has changed in
such a way that your code no longer behaves the same?
Cheers,
David.
On Fri, Sep 23, 2011 at 9:46 AM, Lucian Smith <lpsmith at spod-central.org>
wrote:
>
* Andrew Miller <ak.miller at auckland.ac.nz> [2011-09-22 22:37] writes:
>
> On 23/09/11 06:05, Lucian Smith wrote:
>
> >Out of the approximately 2,000 cellml model files on models.cellml.org,
>
> >over half of them give me an error when I read them in with the CellML
>
> >API, of the form:
>
> >
>
> > ? CellML Variable Po in component TempRLC has in interfaces, but is not
>
> > ? connected (directly or indirectly) to any variable with no in
>
> > ? interfaces;
>
> >
>
> >The thing is, I don't actually care about this error at all--I'm not
>
> >trying to simulate anything, I'm only trying to translate the model, and
>
> >it should be perfectly possible to translate a model like this.
>
>
>
> Hi Lucian,
>
>
>
> Can you remind me why you need to be using CeVAS? The whole point of
>
> CeVAS is to map variables in components to their 'source' (that is, the
>
> variable which is connected to them but is not an in variable).
>
>
>
> Therefore, CeVAS, in its current form, is only useful for variables that
>
> are ultimately connected up to a variable with no in interfaces.
>
>
It looks like I use CeVAS for the function 'iterateRelevantComponents'.
>
>
> This should apply for any complete models, but it won't apply to, say, a
>
> model that is not a complete model, but is designed as a component to be
>
> imported into another model.
>
>
>
> If Antimony can't deal with these component models, then there is little
>
> benefit to trying to translate them. If it can, I'd suggest doing a more
>
> direct translation and letting Antimony work out the connections rather
>
> than using CeVAS to flatten all the parts of the model into one.
>
>
Antimony can indeed deal with thse component models just fine, as it does
>
not have any concept of under-specification of a variable. ?It simply
>
writes out what it has.
>
>
I do indeed work out the connections on my own. I think. ?It's been a
>
while since I wrote that code. ?At the very least, the only times I call
>
'cevas->' anything is to call 'modelError' and
>
'iterateRelevantComponents'.
>
>
-Lucian
>
_______________________________________________
>
cellml-tools-developers mailing list
>
cellml-tools-developers at cellml.org
>
http://lists.cellml.org/mailman/listinfo/cellml-tools-developers
>
Archive powered by MHonArc 2.6.18.