DOC: Best randomness with Parallel generators

[paquiteau]

Hi,

I found a small imprecision in the parallel random state example ¹, as using iterative seed is not recommended as per ². This PR update the example to the best practice, and add a reference to the Numpy documentation for the curious reader.

Footnotes

1. https://joblib.readthedocs.io/en/latest/auto_examples/parallel_random_state.html#sphx-glr-auto-examples-parallel-random-state-py ↩

2. https://numpy.org/doc/stable/reference/random/parallel.html#sequence-of-integer-seeds ↩

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (0c57c18) 94.82% compared to head (c0bb184) 94.84%.
Report is 17 commits behind head on master.

❗ Current head c0bb184 differs from pull request most recent head 32f3ea7. Consider uploading reports for the commit 32f3ea7 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1490      +/-   ##
==========================================
+ Coverage   94.82%   94.84%   +0.02%     
==========================================
  Files          45       45              
  Lines        7491     7491              
==========================================
+ Hits         7103     7105       +2     
+ Misses        388      386       -2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[ogrisel]

The deadlock in the shutdown of the linux_py310_distributed tests is unrelated.

The read the docs failure seems unrelated as well:

Problem in your project's configuration. Invalid configuration option "build.os": build not found.

but it's unfortunate because I would have liked the updated example in this PR to be executed on readthedocs before merging.

[ogrisel]

It did run successfully in the previous iteration though:

https://readthedocs.org/projects/joblib/builds/22563643/

Let's merge.

[ogrisel]

Thanks @paquiteau!