Fixing keyword args in top-level pm.compile_fn

[dfm]

Model.compile_fn requires all parameters except outs to be keyword arguments so the top level compile_fn didn't work as written.

Thank your for opening a PR!

Before you proceed, please make sure that the pre-commit linting/style checks pass.

Depending on what your PR does, here are a few things you might want to address in the description:

• what are the (breaking) changes that this PR makes?
• important background, or details about the implementation
• are the changes—especially new features—covered by tests and docstrings?
• consider adding/updating relevant example notebooks
• right before it's ready to merge, mention the PR in the RELEASE-NOTES.md

[codecov]

Codecov Report

Merging #5865 (84dc7b4) into main (575454f) will decrease coverage by 3.70%.
The diff coverage is 0.00%.

@@            Coverage Diff             @@
##             main    #5865      +/-   ##
==========================================
- Coverage   89.48%   85.78%   -3.71%     
==========================================
  Files          73       73              
  Lines       13225    13225              
==========================================
- Hits        11835    11345     -490     
- Misses       1390     1880     +490     

Impacted Files	Coverage Δ
pymc/model.py	76.42% <0.00%> (-10.85%)	⬇️
pymc/model_graph.py	17.60% <0.00%> (-76.00%)	⬇️
pymc/variational/updates.py	37.93% <0.00%> (-54.19%)	⬇️
pymc/distributions/timeseries.py	43.36% <0.00%> (-35.28%)	⬇️
pymc/data.py	70.20% <0.00%> (-11.43%)	⬇️
pymc/step_methods/hmc/quadpotential.py	73.56% <0.00%> (-6.99%)	⬇️
pymc/backends/arviz.py	85.42% <0.00%> (-4.86%)	⬇️
pymc/printing.py	21.90% <0.00%> (-4.77%)	⬇️
pymc/util.py	73.65% <0.00%> (-3.60%)	⬇️
... and 5 more

[twiecki]

Maybe add a test?

[dfm]

This wasn't previously tested (hence why the bug existed!) so I'd say that that's beyond the scope of my PR, but if someone else wants to, that would be cool.

[ricardoV94]

This wasn't previously tested (hence why the bug existed!) so I'd say that that's beyond the scope of my PR, but if someone else wants to, that would be cool.

Yeah we should test, so we weren't calling this function anywhere? It doesn't need to be you if you don't want to ofc. Thanks for opening the PR

[ricardoV94]

Thanks @dfm!

[dfm]

Wait until we see if the test executes - I just wrote it on the GitHub UI :D

[ricardoV94]

I don't think that should be the case :P (that it only accepts one output). Did we introduce that constraint somewhere by accident?

[dfm]

I'm not sure - it's still not working with a strange error message!

Stale

[dfm]

I'm not sure why my proposed test isn't working, but I'll close this PR and just open an issue and someone else can deal with it because I don't have time!

My proposed 5 character change fixes things in my case but YMMV :D