Fix OpenEphys interface issue with multiple probes of which some are disabled

[vigji]

PR addressing this issue: SpikeInterface/spikeinterface#3539.

Very minimal fix, simply filtering the XML nodes corresponding to enabled probes.

Not sure if this is to be tested? I assume there would be the need for custom asset data, have not looked deeply into it.

[alejoe91]

Thanks @vigji

If you have a settings file with disabled probes it would be great if you added it to the test/data/openephys and also make small test for it :)

[alejoe91]

Note that tests are failing becuase probably the isEnabled has been added with v0.7 of the OE NPIX plugin. Can you fix?

[vigji]

Indeed, I just checked data from the old setup and indeed it is a 0.7 new feature!

yess, I’ll fix this and add a test.

[vigji]

Ok, there should now be new assets and tests for multiple conditions of enabled/disable probes. Let me know how this looks!

[codecov]

Codecov Report

Attention: Patch coverage is 50.00000% with 3 lines in your changes missing coverage. Please review.

Project coverage is 89.42%. Comparing base (1fc2d35) to head (705c0d1).
Report is 52 commits behind head on main.

Files with missing lines	Patch %	Lines
src/probeinterface/io.py	50.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #308      +/-   ##
==========================================
+ Coverage   89.38%   89.42%   +0.03%     
==========================================
  Files          10       10              
  Lines        1885     1891       +6     
==========================================
+ Hits         1685     1691       +6     
  Misses        200      200              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[alejoe91]

Can you run pre-commit locally? The automatic pre-commit fails because it cannot push to organizations

[vigji]

Should be good to go now!

[vigji]

Wait before merging, I am having an issue.

I suspected there was something wrong as I expected this behavior to start happening with new version 0.7 of oephys plugin. However, reading the xml file I was finding

<SETTINGS>
  <INFO>
    <VERSION>0.6.7</VERSION>
   ...

So I assumed I was incorrect and that was the version. However, testing the PR on some files recorded with the non-updated system in the lab, I get same version tag and no isEnabled field.

So, it seems that OEphys is writing xml files with the wrong version in them? is this possible?

Definitively to be checked before merging this and causing troubles to anyone recording with 0.6.7!

[alejoe91]

You could also simply check if the isEnabled field is present and use it if it is. This should be ok for all versions

[vigji]

That's what I am doing now locally but I liked the idea of keeping the logic related explicitly to versions. If that is ok I'll push that - although I still want to understand what is happening with the xml files, I opened an issue for that.

[alejoe91]

@vigji see my last comment

[vigji]

I actually fixed the issue with version check, which as I mentioned I find more explicit and nicer, if you agree I'll leave it this way!

[alejoe91]

Thanks @vigji for this great addition! Merging :)