Add warning if observed in DensityDist is dict

[symeneses]

What is this PR about?
It adds a warning if the observed parameter is a dictionary.

Checklist

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

Major / Breaking Changes

• ...

Bugfixes / New features

• ...
  It solves Raise informative error if dictionary is passed as observed of DensityDist  #6032

Docs / Maintenance

• ...

[codecov]

Codecov Report

Merging #6292 (e933fea) into main (5d7283e) will increase coverage by 0.05%.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #6292      +/-   ##
==========================================
+ Coverage   94.12%   94.17%   +0.05%     
==========================================
  Files         111      111              
  Lines       23861    23921      +60     
==========================================
+ Hits        22458    22528      +70     
+ Misses       1403     1393      -10     

Impacted Files	Coverage Δ
pymc/distributions/distribution.py	95.06% <100.00%> (+0.04%)	⬆️
pymc/tests/distributions/test_distribution.py	98.21% <100.00%> (+0.04%)	⬆️
pymc/step_methods/hmc/quadpotential.py	80.62% <0.00%> (-0.07%)	⬇️
pymc/distributions/discrete.py	99.22% <0.00%> (ø)
pymc/tests/sampling/test_jax.py	100.00% <0.00%> (ø)
pymc/distributions/continuous.py	97.50% <0.00%> (ø)
pymc/variational/opvi.py	87.08% <0.00%> (+0.02%)	⬆️
pymc/step_methods/hmc/nuts.py	97.31% <0.00%> (+0.05%)	⬆️
pymc/sampling/jax.py	97.34% <0.00%> (+0.14%)	⬆️
pymc/variational/callbacks.py	96.22% <0.00%> (+0.22%)	⬆️
... and 7 more

[fonnesbeck]

LGTM. Thanks for contributing!

[michaelosthege]

just some code style simplifications/FYI

looks good!

[ricardoV94]

This is a great start but misses the most important point of the original issue.

DensityDist is defined differently than how it was in V3 where a dictionary with the other parameters besides data could be used. The error message should address this change in API more than what types of observed data are allowed (that's not what users find confusing).

[symeneses]

Based on this item in the Release notes:

pm.DensityDist now accepts distribution parameters as positional arguments. Passing them as a dictionary in the observed keyword argument is no longer supported and will raise an error (see 5026).

I am suggesting this new error message:

Since v4.0.0 the observed parameter should be of type pd.Series, np.array, or pm.Data.
Previous versions allowed passing distribution parameters as a dictionary in observed, in the current version these parameters are positional arguments.

Let me know if this message can be further improved.

[ricardoV94]

@symeneses that sounds good