Add the ability to draw hypergraphs as a bipartite graph

[nwlandry]

This PR does the following:

• Removes draw_dihypergraph and moves its contents to draw_bipartite.
• The draw_bipartite now handles Hypergraph and DiHypergraph objects.
• Added the bipartite_spring_layout position function.

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 91.12426% with 15 lines in your changes are missing coverage. Please review.

Project coverage is 92.29%. Comparing base (0f9be8d) to head (730bd74).

Files	Patch %	Lines
xgi/drawing/draw.py	91.33%	13 Missing :warning:
xgi/drawing/layout.py	88.88%	2 Missing :warning:

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #492      +/-   ##
==========================================
+ Coverage   92.16%   92.29%   +0.13%     
==========================================
  Files          60       60              
  Lines        4418     4493      +75     
==========================================
+ Hits         4072     4147      +75     
  Misses        346      346              

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[maximelucas]

Just a thought: conceptually, could this function just be the same as the draw_dihypergraph with the edge markers, except that the links are drawn without arrows and the layout would be different?

[nwlandry]

Mmmmm. This is a great point. I think you're right. Would it make sense to group them together, call it draw_bipartite and have different handling for directed and undirected hypergraphs?

[maximelucas]

I thought that could be a possibility. Worth thinking about. The current dihypergraph one only make sense as a bipartite if the edge markers are drawn though, I'd say.

Maybe directedness and "bipartite-ness" are two somewhat independent elements here. Could the arrow drawing be included as a directed option in draw_hyperedges? Not sure that would include the edge markers.

And then the draw_bipartite could just be about the layout really. Either the current layout we have in the dihypergraph one, or a more bipartitey layout, the user could choose. And draw directed or undirected edges.

[nwlandry]

I thought that could be a possibility. Worth thinking about. The current dihypergraph one only make sense as a bipartite if the edge markers are drawn though, I'd say.

Maybe directedness and "bipartite-ness" are two somewhat independent elements here. Could the arrow drawing be included as a directed option in draw_hyperedges? Not sure that would include the edge markers.

And then the draw_bipartite could just be about the layout really. Either the current layout we have in the dihypergraph one, or a more bipartitey layout, the user could choose. And draw directed or undirected edges.

I have refactored the function so that it accepts both directed and undirected hypergraphs. Because of this, draw_dihypergraph is now unnecessary and so I removed it.

[review-notebook-app[bot]]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[nwlandry]

I have now implemented all of your suggestions, @maximelucas. Should be ready to re-review.

[nwlandry]

Thanks for the suggestions, @maximelucas! It looks like I need to take more time to address these concerns, and I'll let you know when I'm done!

[nwlandry]

@maximelucas, I believe I addressed all of your comments! Thanks for such a thorough and helpful review! I did the following:

• I standardized all of the function arguments in response to your review.
• I made draw_undirected_dyads and draw_directed_dyads standalone functions and made them available to the user.
• Added default values to the settings and made them match other functions.

[maximelucas]

Thanks! Those are quick small comments for the last changes, I'll have a better look at the entire code soon.

[nwlandry]

@maximelucas, are you still able to review this? Thanks so much!

[maximelucas]

I'll do that next week!

[nwlandry]

Great! Thanks so much!

[maximelucas]

Some more minor comments.. we're almost there, it's a lot of code 😄 I'll check the rest of the code later today.

[nwlandry]

@maximelucas thanks so much! Let me know when you're done and I'll start addressing your comments!

[maximelucas]

I think I checked all of it 😛 I would just add an example somewhere of an undirected one, and raise an issue to create a new function with the "classic" graph bipartite layout, and maybe also with the barycenter layout. Thanks for all the work!

[nwlandry]

@maximelucas, thanks for the review! A quick overview of my changes based on your comments:

• changed all of the dyads to be colored by the edge size (with "crest_r") by default
• Added error handling in a few places
• Added a new function edge_positions_from_barycenters as a more general replacement of the old functionality.
• Added an example of using draw_bipartite on an undirected hypergraph in Tutorial 5
• Added a few examples of the new edge_positions_from_barycenters in Tutorial 5 and In Depth 3.

At this point, I believe that I've addressed all your comments. Let me know if there are other things you would like me to change.

[maximelucas]

Looks good, thanks so much for the work!