Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

leiden-modularity-plugin: Crashes when filters are active? #216

Open
deklanw opened this issue Jan 29, 2020 · 4 comments
Open

leiden-modularity-plugin: Crashes when filters are active? #216

deklanw opened this issue Jan 29, 2020 · 4 comments

Comments

@deklanw
Copy link

deklanw commented Jan 29, 2020

@vtraag

Environment: Windows. v0.9.2.

The plugin works fine when I haven't applied any filters. If I apply a filter and run it it stops immediately, there is no popup window, and I see Leiden algorithm 0 in the panel on the right. If I try to run it again I see a spinner and the program becomes unresponsive. I have to close with Task Manager

@vtraag
Copy link

vtraag commented Jan 29, 2020

Thanks for reporting the issue! I will look into it.

@eduramiba, feel free to assign me the issue (if that is possible).

@deklanw
Copy link
Author

deklanw commented Jan 29, 2020

@vtraag Awesome, thanks for the quick reply.

While I have your attention... I have a question (I'm an amateur at this). It seems that at a certain point of resolution and lower I just get 1 cluster. Why would that be?

Example of some test runs of a graph with 1.2k nodes and 69k edges, 1000 iterations, 20 restarts.

Resolution Clusters Quality
0.5 217 0.072
0.4 163 0.110
0.3 125 0.160
0.2 89 0.232 
0.15 72 0.285
0.13 65 0.310
0.12 52 0.324
0.11 61 0.338
0.1 55 0.354 
0.095 61 0.362
0.09 62 0.371
0.085 53 0.380
0.08 53 0.389
0.07 55 0.414
0.065 52 0.427
0.06 50 0.442
0.055 46 0.458
0.05 41 0.479 (too much one cluster starts here and below)
0.04 46 0.539
0.03 49 0.612
0.02 33 0.711 (basically all one cluster)
0.01 1 0.845

Starting at about 0.05 resolution, a single cluster dominates almost the entire graph (despite there being 41 clusters). Eventually, at 0.01 there is a single cluster

@vtraag
Copy link

vtraag commented Jan 29, 2020

This is normal behaviour. For a certain resolution parameter lower than some threshold, you will always find a single cluster.

This is probably not the most relevant clustering, so you might want to use one of the higher resolution parameters. What resolution parameter(s) is/are best suited for your particular analysis depends on the context and your needs. Feel free to simply mail me, then we don't clutter this issue.

@mbastian
Copy link
Member

Hi @vtraag I can also reproduce this issue. I've assigned it to you but I'm happy to assist if needed.

This is the exception I received

java.lang.IllegalArgumentException: Last element of firstNeighborIndices array must be equal to nEdges.
	at cwts.networkanalysis.Network.checkIntegrity(Network.java:1199)
	at cwts.networkanalysis.Network.<init>(Network.java:1374)
	at cwts.networkanalysis.Network.<init>(Network.java:275)
	at cwts.networkanalysis.gephiplugin.RunClustering.execute(RunClustering.java:111)
	at org.gephi.statistics.StatisticsControllerImpl.execute(StatisticsControllerImpl.java:163)
	at org.gephi.statistics.StatisticsControllerImpl$3.run(StatisticsControllerImpl.java:150)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

3 participants