documentation: workflow, badges, workshop link, ftest warnings

[cecileane]

No description provided.

[cecileane]

@pbastide : if you could look at the explanations for the warnings sent by GLM.ftest in the documentation, that would be nice. These explanations are in an "admonition" box at the end of the page on continuous traits, previewed here. But if you don't have time, let me know, that's fine. It should not hurt to add documentation!

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 86.74%. Comparing base (787582f) to head (c1d37b8).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #212      +/-   ##
==========================================
- Coverage   86.80%   86.74%   -0.06%     
==========================================
  Files          31       31              
  Lines       12638    12637       -1     
==========================================
- Hits        10970    10962       -8     
- Misses       1668     1675       +7     

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

[pbastide]

@pbastide : if you could look at the explanations for the warnings sent by GLM.ftest in the documentation, that would be nice. These explanations are in an "admonition" box at the end of the page on continuous traits, previewed here. But if you don't have time, let me know, that's fine. It should not hurt to add documentation!

Yes, the explanations make sense, thank you !
It would be easy to correct the r2 values in the table : see comment in the code.
Then we could suppress the warnings altogether ? But I'm not sure how to do that properly in julia.

[cecileane]

We could send logging at the warning level to devnull to suppress the warning, but

• that would have the downside of adding one dependency (Logging)
• I don't know how to filter specific warnings to make sure that only this specific warning is suppressed. If we wanted to suppress logging information at the warning level only and not at the error level (from @error, which does not throw an error like error) I would have to learn how to do that.
• more importantly, I feel uncomfortable suppressing logging that's not specific.

I agree with your idea that redefining ftest is the best way to go... either in GLM via a pull request, or here specifically (not ideal, but less work). Right now I have no time for this unfortunately :(

[pbastide]

I agree that suppressing all warnings without being specific is not satisfying, and I could not find an easy way to do this.
I am good with the note that you wrote to explain the warnings, until we find a better way.
Thanks !