Rebuild for arpack 3.9

[regro-cf-autotick-bot]

This PR has been triggered in an effort to update arpack39.

Notes and instructions for merging this PR:

1. Please merge the PR only after the tests have passed.
2. Feel free to push to the bot's branch to update this PR if needed.

Please note that if you close this PR we presume that the feedstock has been rebuilt, so if you are going to perform the rebuild yourself don't close this PR until the your rebuild has been merged.

---

If this PR was opened in error or needs to be updated please add the bot-rerun label to this PR. The bot will close this PR and schedule another one. If you do not have permissions to add this label, you can use the phrase code>@<space/conda-forge-admin, please rerun bot in a PR comment to have the conda-forge-admin add it for you.

_{This PR was created by the regro-cf-autotick-bot. The regro-cf-autotick-bot is a service to automatically track the dependency graph, migrate packages, and propose package version updates for conda-forge. Feel free to drop us a line if there are any issues! This PR was generated by https://github.com/regro/cf-scripts/actions/runs/9263225890 - please use this URL for debugging.}

[conda-forge-webservices[bot]]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[xylar]

The errors are:

======================================================================
ERROR: testAuthorityScore (tests.test_structural.CentralityTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/conda/feedstock_root/build_artifacts/python-igraph_1716879267785/test_tmp/tests/test_structural.py", line 499, in testAuthorityScore
    asc = g.authority_score()
igraph._igraph.InternalError: Error at src/linalg/arpack.c:1098: ARPACK error. -- No shifts could be applied during a cycle of the Implicitly restarted Arnoldi iteration. One possibility is to increase the size of NCV relative to NEV

======================================================================
FAIL: testHubScore (tests.test_atlas.GraphAtlasTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/conda/feedstock_root/build_artifacts/python-igraph_1716879267785/test_tmp/tests/test_atlas.py", line 103, in testHubScore
    sc = g.hub_score()
igraph._igraph.InternalError: Error at src/linalg/arpack.c:1098: ARPACK error. -- No shifts could be applied during a cycle of the Implicitly restarted Arnoldi iteration. One possibility is to increase the size of NCV relative to NEV

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/conda/feedstock_root/build_artifacts/python-igraph_1716879267785/test_tmp/tests/test_atlas.py", line 105, in testHubScore
    self.assertTrue(
AssertionError: False is not true : Hub score calculation threw exception for graph #224: Error at src/linalg/arpack.c:1098: ARPACK error. -- No shifts could be applied during a cycle of the Implicitly restarted Arnoldi iteration. One possibility is to increase the size of NCV relative to NEV

@vtraag, any insight?

[vtraag]

No, I don't have any additional insight at the moment. It seems that this remains a recurrent issue for which it is unclear how to resolve it.

[xylar]

Is this the same error we were getting intermittently that caused us to disable these tests before?

[vtraag]

Yes, re-disabling the tests should do the trick. Still not clear why these fail though.

[xylar]

Okay, I'll rebase and remove my commits that re-enabled the tests.

[szhorvat]

AssertionError: False is not true : Hub score calculation threw exception for graph #224: Error at src/linalg/arpack.c:1098: ARPACK error. -- No shifts could be applied during a cycle of the Implicitly restarted Arnoldi iteration. One possibility is to increase the size of NCV relative to NEV

Can you please re-run these tests and see if the failure occurs for the same graph 224?

I'll have a look if there's any manipulation of the starting vector that we can do to improve reliability, but no promises ... (I'm not hopeful).

It's frustrating that I can't repro the issue locally. If we could find a relatively quick turnaround way to test changes, that would be very helpful.

[xylar]

@szhorvat, I removed the patch again that disables the tests. Let's see what the failures look like this time.

[xylar]

As we have seen in the past, the errors seem different each time. Here's what I'm seeing now (for linux 64 python 3.9):

======================================================================
ERROR: testHubScore (tests.test_structural.CentralityTests.testHubScore)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/conda/feedstock_root/build_artifacts/python-igraph_1716906509993/test_tmp/tests/test_structural.py", line 511, in testHubScore
    g.hub_score(scale=False, return_eigenvalue=True)
igraph._igraph.InternalError: Error at src/linalg/arpack.c:1098: ARPACK error. -- No shifts could be applied during a cycle of the Implicitly restarted Arnoldi iteration. One possibility is to increase the size of NCV relative to NEV

----------------------------------------------------------------------
Ran 453 tests in 2.678s

[xylar]

And for python 3.12:

======================================================================
ERROR: testHubScore (tests.test_structural.CentralityTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/conda/feedstock_root/build_artifacts/python-igraph_1716906537148/test_tmp/tests/test_structural.py", line 511, in testHubScore
    g.hub_score(scale=False, return_eigenvalue=True)
igraph._igraph.InternalError: Error at src/linalg/arpack.c:1098: ARPACK error. -- No shifts could be applied during a cycle of the Implicitly restarted Arnoldi iteration. One possibility is to increase the size of NCV relative to NEV

----------------------------------------------------------------------
Ran 453 tests in 3.115s

[szhorvat]

Thank you!

Let's leave it at this for now.

I have some ideas for improving robustness. They're based on pure guesswork, but it's worth a try. I'll implement them and let's come back to this issue when the next igraph version is released.

[xylar]

Okay, thanks @szhorvat! I'll disable the failing tests again for now and we'll try again soon!

[szhorvat]

I made some changes, some of which will be in the next version of C/igraph (0.10.13), and some only in the 1.0 release.

Unfortunately I don't think this issue can be fully fixed. ARPACK is fickle by nature, and there will always cases where it fails with any specific choice of default parameters. I think the test that's failing here is a bit too eager.

[minrk]

I'm checking up on various conda-forge migrations. I'm not sure I follow the discussion above, is this one ready to merge?

[xylar]

@minrk, apologies. Yes, I think it was ready to merge.