Extend Channel fields to include URL to NeuroML file

[slarson]

We want bioparameters associated with ion channels as well as metadata indicating level of certainty to appear in PyOpenWorm so that they can be read in the dashboard.

Examples of bioparameters being built into a model can be found here and here

'E_rev', cond_density (g_des), v_half_a, k_a, T_a, should be the initial parameters, with two metadata parameters for each.

Need to reconcile the metadata parameters from NeuroML with the PyOpenWorm concept of Evidence. For now, let's use the Evidence mechanism as the way of incorporating the two metadata parameters into the Ion Channel. We may need help from @mwatts15 to understand how we can use evidence in a per-parameter way to achieve this.

One example of how this might be is E_rev's value of -70 mV is at 'BlindGuess' and 0.1 certainty, while v_half_a is at 30 mV with 'BlindGuess' at 0.3 certainty. We want the evidence to tell us that the v_half_a value came from running the ChannelWorm fitter and the E_rev came from Stephen Larson's personal best guess. Other values could come from published papers with DOIs.

[slarson]

@shubhsingh594 Any luck with this so far? Talking to @VahidGh today, it may be better for us to be storing the NeuroML file that represents the ion channel model directly, rather than breaking it down into its separate parameters and storing each of them explicitly in PyOpenWorm. Main reason is that different ion channel files are likely to have different numbers of relevant parameters and may be a pain for data representation. Let me know what you think about this?

[shubhsingh594]

@slarson , so should we just have link to the corresponding NeuroML files?
or at some point we should parse the file, get the available relevant data and show that on the dashboard.

[slarson]

Let's settle on importing the full text of the channel.nml file, along with the metadata parameters into PyOpenWorm. Then we'll have a reader parse the channel.nml file on the output side. This is going to be easier because the data model for a channel.nml file is complex and we don't want to reproduce it wholesale in PyOpenWorm.

[slarson]

@tarelli Made a good suggestion -- can we point at a URL with the nml files instead of the full text?

[slarson]

Update on this -- not extending the field to include the data, just the URL.