Pandas simplification v2

[CalCraven]

This is a replacement PR for the features requested in #814, which has some commits in the merge history that deviated when switching to ruff linting.

[CalCraven]

From previous PR:

This PR looks to improve the handling for converting a topology to a dataframe. This currently lives as a method for topology. It is now being moved to a convert_dataframe.py module. A few different formats are available which give some nice default ways to view a topology. Notably, we have the formats:
-publication which gives all the parameter values you would want to have in a table for publication. This also removes duplicates so each parameter is only listed once.
-default some default values which are nice to have
-remove_duplicates which allows you to get a smaller dataframe with duplicate rows removed.
-specific_columns Allows the user to specify what they want in the dataframe.

There is also an added function that allows you to generate dataframes that cover the parameters for a set of topologies.

Finally, there will be some function that prints the dataframes with the rdkit mols which are labeled to match the dataframes.

TODO Checklist:

• Error checking on arguments
• Replace topology.py dataframe methods/tests
• Doc strings
• Handle units
• Handle parameters that return lists
• Handle parameters that return dictionaries
• Return unique elements if style is publication
• Handle parameter "all" better
• Function to concatenate multiple topogies into one output
• Remove replicate rows flag -> similar to the publication style, but without the atom_indices section added
• Function to create a topology with all data from dataframe matching rdkit mol image
  • Could just also make this a format unique_types

[CalCraven]

From discussion with @Vtsoch, there are a few more general use cases that would be nice to have working in the arguments for the main function, to_dataframeDict.

dfDict = to_dataframeDict(ptop, parameters="sites", columns=["name", "atom_type.name", "atom_type.parameters", "charge", "molecule.name"], format="remove_duplicates")

should be put as an example, since it could be hard to find the molecule info or parameters info if you don't know how the parsing works of these attributes. I would even consider these attributes to be in the default format since they're nice to know.

dfDict = to_dataframeDict(ptop, parameters=["sites", "bonds"], columns=["name"], format="specific_columns")

This will fail currently. However, I think it should just go through the columns and only grab attributes that exist, skipping the others.

[codecov]

Codecov Report

Attention: Patch coverage is 91.66667% with 13 lines in your changes missing coverage. Please review.

Project coverage is 93.31%. Comparing base (5a4f17d) to head (72037d5).
Report is 6 commits behind head on main.

Files	Patch %	Lines
gmso/external/convert_dataframe.py	91.55%	13 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #826      +/-   ##
==========================================
- Coverage   94.07%   93.31%   -0.77%     
==========================================
  Files          65       66       +1     
  Lines        6953     7088     +135     
==========================================
+ Hits         6541     6614      +73     
- Misses        412      474      +62     

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

[chrisjonesBSU]

Left a couple comments.

[chrisjonesBSU]

This should be dict correct?

[chrisjonesBSU]

@Zeerakkhan47 I think you could expand the information passed into the columns parameter here to include some of the things @CalCraven mention in his comment on June 16th.