Add a planarity notebook #146

ayra0812 · 2025-03-28T11:24:20Z

I recently submitted a pull request for the Planarity Notebook in the NetworkX repository. I noticed that there are some minor errors in my submission, and I sincerely appreciate your time in reviewing my work.

I am eager to contribute to this project and improve my submission to align with the repository's standards. Please let me know the necessary changes or improvements you'd like me to make, and I'll be happy to update the PR accordingly.

Looking forward to your feedback. Thank you for your guidance and for this opportunity—I’m truly excited to contribute and learn.

Fixes #122

dschult · 2025-04-18T17:28:29Z

It looks like this PR has a few extra files from isomorphism. These are probably additions you were making to update the Pr to match the main branch. I think it is probably best to "merge" or "rebase" the main branch changes into your branch rather than add them manually through commits and patches to the PR.

To revert the recent commits that caused the problem you can git reset --hard <place to reset HEAD to>. I usually create another branch at the current head in case something goes wrong: git checkout -b new_name then git checkout old_name then git reset....

Then to rebase/merge with the main branch

make sure you update your local main branch to by pulling from the nx-guides repo.
checkout the PR branch
git rebase -i main or if that gets too ugly (too many commits to keep track of which changes go into which commit) then try git merge main.
- Any conflicting lines will stop the rebase process and git will tell you which files have conflicts. You can then edit the files where the conflicts are marked (search for HEAD). Then you commit those changes and use git rebase --continue to continue to the next set of changes. This is sometimes an onerous task. Ask for help. A merge is easier in the sense that you put all changes together into a single commit so conflicts only have to be managed once for all the commits in your PR. But rebase is better in the sense that the resulting PR has all changes in the most recent commits (merge spreads them out between other commits).
- You can stop the merge/rebase if you want to start over by using the command git rebase --abort or git merge --abort. The git messages tell you about this too.
After the rebase or merge successfully finishes you can then push it to your github fork and it automatically updates the PR. This push to your repo will require a git push --force option because it is both removing and adding commits to the remote branch.

There may be places on the web that describe merge/rebase better than I have here. If you have troubles post here and we'll try to figure it out too.

ayra0812 · 2025-04-19T05:33:53Z

It looks like this PR has a few extra files from isomorphism. These are probably additions you were making to update the Pr to match the main branch. I think it is probably best to "merge" or "rebase" the main branch changes into your branch rather than add them manually through commits and patches to the PR.

To revert the recent commits that caused the problem you can git reset --hard <place to reset HEAD to>. I usually create another branch at the current head in case something goes wrong: git checkout -b new_name then git checkout old_name then git reset....

Then to rebase/merge with the main branch

make sure you update your local main branch to by pulling from the nx-guides repo.

checkout the PR branch

git rebase -i main or if that gets too ugly (too many commits to keep track of which changes go into which commit) then try git merge main.

Any conflicting lines will stop the rebase process and git will tell you which files have conflicts. You can then edit the files where the conflicts are marked (search for HEAD). Then you commit those changes and use git rebase --continue to continue to the next set of changes. This is sometimes an onerous task. Ask for help. A merge is easier in the sense that you put all changes together into a single commit so conflicts only have to be managed once for all the commits in your PR. But rebase is better in the sense that the resulting PR has all changes in the most recent commits (merge spreads them out between other commits).

You can stop the merge/rebase if you want to start over by using the command git rebase --abort or git merge --abort. The git messages tell you about this too.

After the rebase or merge successfully finishes you can then push it to your github fork and it automatically updates the PR. This push to your repo will require a git push --force option because it is both removing and adding commits to the remote branch.

There may be places on the web that describe merge/rebase better than I have here. If you have troubles post here and we'll try to figure it out too.

Yes, I followed the same procedure by resetting all the previous changes. After adding the planarity section, I merged it with the main branch to retrieve the remaining files. Now, I am at this stage. Please take a moment to review it.

dschult · 2025-04-14T21:59:49Z

content/algorithms/planarity/planarity.md

+
+G: NetworkX graph
+counterexample : bool
+A Kuratowski subgraph (to proof non planarity) is only returned if set to true.


Introduce the term "Kuratowski subgraph". Each time a technical term is introduced say what it is. In this case, perhaps you can say this when you introduce the theorem above.

typo?

Suggested change

A Kuratowski subgraph (to proof non planarity) is only returned if set to true.

A Kuratowski subgraph (to prove non planarity) is only returned if set to true.

dschult · 2025-04-30T16:32:10Z

content/algorithms/planarity/planarity.md

+## Planarity Algorithm in NetworkX
+
+## Introduction
+A graph is said to be **planar** if it can be drawn on a plane without any of its edges crossing. The `check_planarity` function in NetworkX helps determine whether a given graph is planar and provides either an embedding or a counterexample.


Maybe explain what an embedding is. And what we mean by a counterexample.

dschult · 2025-04-30T16:32:58Z

content/algorithms/planarity/planarity.md

+### Mathematical Background
+A graph is **planar** if and only if it does not contain a subgraph homeomorphic to **K₅ (complete graph on 5 vertices)** or **K₃,₃ (complete bipartite graph with partition sizes 3 and 3)**. This is known as **Kuratowski’s theorem**.
+
+The planarity check in NetworkX is based on the **Left-Right Planarity Test**, an efficient combinatorial method to determine planarity.


Can you describe briefly in a few sentences what the left-right planarity test does?

Sure Sir, I will work on all the suggestions and make my Jupyter notebook more understandable.

dschult

This guide includes a lot of code, and some text. But the connection between the code and the text isn’t always clear, and the code is fairly hard to read. Can you run the NetworkX code but explain what it is doing in words — rather than making the reader read the code?

It would also be good to include more about why one would want todo this and why it is an important idea/question/problem.

dschult · 2025-05-16T05:03:30Z

content/algorithms/planarity/planarity.md

+```python
+#Example 1
+# Load GraphML file
+G = nx.read_graphml("./data/planar_graph.graphml")


You can create the graph in code for the small graphs in this notebook. That is more readable because they can see the nodes and edges without all the code and distraction of loading the file.

G = nx.Graph([(0, 1), (1, 2), (2, 3), (0, 2)])

dschult · 2025-05-16T05:06:21Z

content/algorithms/planarity/planarity.md

+
+
+```python
+from functools import singledispatch


This line of code is not needed/used in the notebook. Try to remove any spurious cells and code. :)

dschult · 2025-05-16T05:07:45Z

content/algorithms/planarity/planarity.md

+
+
+```python
+def is_planar(G):


Are these functions part of NetworkX? If so, you don’t need to define them here. You can call them from e.g. nx.is_planar

dschult · 2025-05-16T05:10:50Z

content/algorithms/planarity/planarity.md

+
+## Notes
+
+**Embedding** - A (combinatorial) embedding consists of cyclic orderings of the incident edges at each vertex. Given such an embedding there are multiple approaches discussed in literature to drawing the graph (subject to various constraints, e.g. integer coordinates), see e.g. [2].


I think an embedding is an assignment of coordinates in the plane to each node. The cyclic ordering of edges can be built as a way to find an embedding. Does combinatorial embedding mean the cyclic ordering even though it doesn’t given the coordinates? It is possible I am missing that key point. :)

dschult · 2025-05-16T05:14:16Z

content/algorithms/planarity/planarity.md

+
+
+
+A counterexample is only generated if the corresponding parameter is set, because the complexity of the counterexample generation is higher.


Even if the complexity was the same, it would take significant time so the reason for making it optional is really about it taking time to do, the complexity isn’t as important as the extra time it takes. (which is related to the complexity, but those “constants” in the complexity description mean we can’t base time on complexity directly.

dschult · 2025-05-16T05:16:24Z

content/algorithms/planarity/planarity.md

+
+
+```python
+class PlanarEmbedding(nx.DiGraph):


Including this code makes the notebook hard to read. It is code in NetworkX, so you can call it from that. And it would be good to explain it rather than making the reader figure out how the code works.

dschult changed the title ~~Issue #122~~ Add a planarity notebook Apr 4, 2025

ayra0812 added a commit to ayra0812/nx-guides that referenced this pull request Apr 8, 2025

Add planarity/planarity to TOC in index.md for PR networkx#146

7219555

ayra0812 force-pushed the master branch from 81b24c7 to 22bdf5c Compare April 14, 2025 05:32

Add Planarity notebook

74ae6de

ayra0812 force-pushed the master branch from d477194 to 74ae6de Compare April 19, 2025 04:30

Update index.md

19a7d79

Update index.md

0f31db2

dschult reviewed Apr 30, 2025

View reviewed changes

ayra0812 added 2 commits May 14, 2025 14:26

Updated Notebook

58e6b5f

enhanced notebook understandability

3be83fc

dschult reviewed May 16, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add a planarity notebook #146

Add a planarity notebook #146

ayra0812 commented Mar 28, 2025 •

edited by dschult

Loading

dschult commented Apr 18, 2025

ayra0812 commented Apr 19, 2025

dschult Apr 14, 2025

dschult Apr 30, 2025

dschult Apr 30, 2025

ayra0812 May 3, 2025

dschult left a comment

dschult May 16, 2025

dschult May 16, 2025

dschult May 16, 2025

dschult May 16, 2025

dschult May 16, 2025

dschult May 16, 2025

	A Kuratowski subgraph (to proof non planarity) is only returned if set to true.
	A Kuratowski subgraph (to prove non planarity) is only returned if set to true.


		## Notes

		Embedding - A (combinatorial) embedding consists of cyclic orderings of the incident edges at each vertex. Given such an embedding there are multiple approaches discussed in literature to drawing the graph (subject to various constraints, e.g. integer coordinates), see e.g. [2].




		A counterexample is only generated if the corresponding parameter is set, because the complexity of the counterexample generation is higher.



		```python
		def is_planar(G):



		```python
		class PlanarEmbedding(nx.DiGraph):

Add a planarity notebook #146

Are you sure you want to change the base?

Add a planarity notebook #146

Conversation

ayra0812 commented Mar 28, 2025 • edited by dschult Loading

dschult commented Apr 18, 2025

ayra0812 commented Apr 19, 2025

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dschult left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ayra0812 commented Mar 28, 2025 •

edited by dschult

Loading