fix: update for network_p2p.

[sachinstl]

Changes in plot, community and centrality using Networkx package.

1. In new version Plot is based on networkx package which helped to remove the errors like zerodivision.
2. Changes in community and centrality plots, newer version supports all the community detections like walk_trap, fastgreedy, etc.
3. Changes in random seeds, color palletes etc.

Summary

This branch

Changes

The changes made in this PR are:

1. Change 1
2. Change 2

Checks

• [ ] All checks pass (run python -m unittest discover -s tests -f locally)
• [ ] Versions are updated in pyproject.toml, setup.py, and docs/conf.py (if applicable)
• [ ] NEWS.md has been updated

Notes

This fixes #

<other things, such as how to incorporate new changes>

[szhorvat]

1. In new version Plot is based on networkx package which helped to remove the errors like zerodivision, This error was frequent in igraph implementation.

In what specific case do you see "division by zero" errors with igraph? If you found a problem with igraph, please open an issue and show a minimal example.

[martinctc]

1. In new version Plot is based on networkx package which helped to remove the errors like zerodivision, This error was frequent in igraph implementation.

In what specific case do you see "division by zero" errors with igraph? If you found a problem with igraph, please open an issue and show a minimal example.

Thanks for flagging @szhorvat. I can confirm that the "division by zero" error occurs when an igraph plot is being generated in a graph with either no edge attributes (or potentially no edges), which happens to a graph when simplify() from igraph is run. This behaviour with simplify() (of seemingly removing too much edge data) did not seem to apply to the R version of igraph, and only the Python version of igraph. In PR #18, this issue was resolved by removing the simplify() step from the Python library, which allows us to preserve functionality without switching over to the networkx implementation. We will look into creating a minimal example to demonstrate this potential bug and submit to igraph. Thanks!

[szhorvat]

Thanks so much for tracking this down @martinctc! Can you have a look at this behaviour, @ntamas?

[ntamas]

Thanks so much for tracking this down @martinctc! Can you have a look at this behaviour, @ntamas?

Unfortunately I can't reproduce this with an empty graph (i.e. a graph with no edges) on my end. A SSCCE would be helpful to fix this on our side if it is indeed a bug.

[martinctc]

Thanks so much for tracking this down @martinctc! Can you have a look at this behaviour, @ntamas?

Unfortunately I can't reproduce this with an empty graph (i.e. a graph with no edges) on my end. A SSCCE would be helpful to fix this on our side if it is indeed a bug.

Ah okay. Will work on producing one and revert - thanks for checking @ntamas !