ABI CellML meeting minutes 2010-03-24

Present: Randall Britten, Dougal Cowan, Catherine Lloyd, Lukas Endler, Tommy Yu, Justin Marsh, Andrew Miller, David "Andre" Nickerson, Alan Garny, Poul Nielsen

Apologies: Peter Hunter

This week's agenda:

1) (Catherine) Model annotation: Can we update the API to handle convenience methods for editing metadata as requested by the saint team?

  • Randall mentioned that this has been discussed between Auckland and the Saint developers via email.
  • Catherine reported that the saint developers would rather not handle raw RDF.
  • Andrew suggested that we could write some Java code to do the RDF work, and provide this code.
  • Andre said that Allyson is right, and that users who are annotating should not need to know anything about RDF [including terminology such as "bag"] in order to annotate.
  • There was a discussion about standardizing the way annotation is done, and that this standardization is required.
  • Alan mentioned that there is an ongoing project in Oxford about annotating cardiac electrophysiological models and that, due to time constraints, the Auckland CellML API will not be used.
  • Andre said that we need to do this as an API service, as requested by the Saint developers.
  • There was an extensive discussion about whether to put this into the API or to provide Java code.
  • Andrew said that a Java shim would be quick to develop if we had the RDF bag convenience utilities to build it on. Justin said that a Java shim would be quick to change if it did not fit the needs of the saint team, and that it could become an API service later. An example of this sort of thing happening is the CeLEDS exporter, which moved from OpenCell code to the API.
  • There was a decision to offer to provide Java code to the Saint developers.

2) CellML API - update

  • Andrew has been helping Lucian with the API.
  • Andrew is working on supporting IDA. He has created a mock-up by mixing in the code currently generated with some hand-written code. IDA's initial value solver requires that we specify which variable haves rate constraints on them, which Andrew is also working on. Andrew noted, however, that in his work with ModML, he found that IDA's initial value solver didn't work as well as using an external Levenberg-Marquardt implementation, and so we may need to use the same approach with the API.
  • Justin has been working on tracker item 2448.

3) OpenCell development - update

  • Justin has implemented copy and paste of the major nodes in the model tree view (Components, variables, reactions, Math and equation elements, units, imports, connections).

   1. OpenCell 0.8 feature planning

  • Randall mentioned that Andrew is already working on IDA support.
  • Andrew suggested that a commitment to finish tree view refactoring would be required if we start this, as time has already been wasted (due to bit-rot from further changes in OpenCell) on this.
  • Andre asked if there are any plans to get people to use OpenCell, or to turn OpenCell into a modelling tool, instead of the CellML editing tool it is currently.
  • There was a discussion about why people choose not to use OpenCell, and how we can get people to use it.
  • There was a discussion about possibly providing better facilities for round-tripping to and from MATLAB, to encourage the creation of models in CellML.
  • Randall said that we might want to look at getting OpenCell's MATLAB export up to the standard provided by COR.
  • Alan suggested that LATEX, TEX and Word export in OpenCell would be very useful.
  • There was a discussion about maths export and display, and whether this might be a good thing to provide as an API service, as it would be useful to have in PMR2.
  • There was a discussion about alternative platforms for OpenCell.
  • Andrew noted that we should see if there are any other projects it is worth merging our code with before we switch platforms, as the effort to merge and the effort to switch platforms would mostly overlap.

  • Justin announced that OpenCell 0.7 and the API 1.7 would be released today.

4) cellml.org - update

  • No update.

5) Repository contributions - update

  • Catherine reported that Geoff Nunns has almost completed Grandi 2010.
  • Catherine has been going through the last of the student's models, updating, pushing and exposing them.
  • Catherine has annotated the Goldbeter 1991 model by hand using BioModels qualifiers and has had some feedback from Sarala and Jonathan Cooper.

6) PMR2 development - update

  • Tommy has been working on a cmgui view to render cmgui models (for FieldML repository) and support for storage of license and display for models.
  • Tommy has updated the code generation, using CeLEDS to generate code in the four supported languages. This works with CellML 1.1 except where models import components stored within the same repository due to a deadlock issue.
  • Tommy is working on profiles to provide different listings and views for models.physiome, models.cellml, etc from the repository.
  • Tommy has tentatively put the next release of PMR2 at end of April, with its testing instance brought live some time mid-April.

7) Specifications

  • ModML: Andrew reported that tracing the location of unit errors in ModML is possible but messy, and is restricted by a limitation hard coded into the interpretation environment.
  • ModML: Andrew is designing a schema for reactions. This requires a lot of thinking about how to implement compartments and the communication between them.

8) Metadata specifications (general, citations, graphing, simulation, other)

  • No update.