Adding function for mutable coords

[Dhruvanshu-Joshi]

What is this PR about?
In this PR , I have made attempts to incorporate suggestions made from 6497.
Coordinate dimensions by default are assumed to be immutable unless they are explicitly mentioned to be mutable in the add_coord function.
In this PR, we aim to register mutable dimension coordinates with the model without having to explicitly set mutable = True in the add_coord method for each item in coords.

Checklist

• Explain important implementation details 👆
• Make sure that the pre-commit linting/style checks pass.
• Link relevant issues (preferably in nice commit messages)
• Are the changes covered by tests and docstrings?
• Fill out the short summary sections 👇

Major / Breaking Changes

• no

New features

• Added another function called add_mutable_coords which registers new coordinates that are mutable by iterating in a for loop and calling self.add_coord(..., mutable=True) for every iteration.

Bugfixes

• no

Documentation

• Documentation has been updated consistent with modifications

Maintenance

• no

[twiecki]

Thanks! We should also change the error message to mention to set mutable=True then.

[twiecki]

And I think we should make it mutable by default.

[Dhruvanshu-Joshi]

How would you like to make it mutable by default?
What I was thinking was to create another function called add_mutable_coord which will be same as add_coord but will assume mutable = True. We can also update add_coord to consider only cases when mutable = False. We will also throw error messages when someone tries to add immutable coordinates using add_mutable_coord and vice-versa. Hence by using add_mutable_coords which will call add_mutable_coord, one can add mutable coordinates to the model context and by using add_coords which iteratively calls add_coord, we can add immutable coordinates to the model context.
Should I proceed with this?

[michaelosthege]

Let's keep it as simple as possible:

• Introduce coords_mutable to the Model.__init__ signature
• Put a two-line for iterator right below the self.add_coords(coords) line

There's no need to create another method. Not counting tests this can be a 3-line change!

I'm also against changing defaults because of possible shape or performance issues.

[Dhruvanshu-Joshi]

Hey @twiecki @michaelosthege when we use coords_mutable.items, we are expecting a none-type object to be a dictionary and hence all of the tests are failing. I suggest that we include this loop in an if block which will check
if coords_mutable is not None : similar to how add_coords does it to avoid this issue. Should I proceed with this?

[michaelosthege]

Hey @twiecki @michaelosthege when we use coords_mutable.items, we are expecting a none-type object to be a dictionary and hence all of the tests are failing. I suggest that we include this loop in an if block which will check if coords_mutable is not None : similar to how add_coords does it to avoid this issue. Should I proceed with this?

sure, it's your PR!

[codecov]

Codecov Report

Merging #6515 (1927320) into main (28bac77) will decrease coverage by 25.22%.
The diff coverage is 33.33%.

Additional details and impacted files

@@             Coverage Diff             @@
##             main    #6515       +/-   ##
===========================================
- Coverage   94.74%   69.52%   -25.22%     
===========================================
  Files         146      146               
  Lines       27807    27810        +3     
===========================================
- Hits        26346    19336     -7010     
- Misses       1461     8474     +7013     

Impacted Files	Coverage Δ
pymc/model.py	89.08% <33.33%> (-0.78%)	⬇️
pymc/tests/sampling/test_forward.py	0.00% <0.00%> (-100.00%)	⬇️
pymc/tests/stats/test_convergence.py	0.00% <0.00%> (-100.00%)	⬇️
pymc/tests/distributions/test_bound.py	0.00% <0.00%> (-100.00%)	⬇️
pymc/tests/sampling/test_population.py	0.00% <0.00%> (-100.00%)	⬇️
pymc/tests/step_methods/test_slicer.py	0.00% <0.00%> (-100.00%)	⬇️
pymc/tests/stats/test_log_likelihood.py	0.00% <0.00%> (-100.00%)	⬇️
pymc/tests/step_methods/hmc/test_nuts.py	0.00% <0.00%> (-100.00%)	⬇️
pymc/tests/step_methods/test_compound.py	0.00% <0.00%> (-100.00%)	⬇️
pymc/tests/distributions/test_censored.py	0.00% <0.00%> (-100.00%)	⬇️
... and 43 more