Fix - Do not split abspath with dot

[VaelK]

Hello,

Following issue #1362, this improvement avoid splitting the absolute path whenever a dot is found in it. Let's say that a folder in this path is named name.folder, then the test test_parallel_call_cached_function_defined_in_jupyter does not pass. Plus it create an extra level in the tree structure of backend store.

Let me know if there is a problem with that.

Thank you!

[codecov]

Codecov Report

Base: 93.96% // Head: 93.95% // Decreases project coverage by -0.01% ⚠️

Coverage data is based on head (0de702a) compared to base (796e7aa).
Patch coverage: 100.00% of modified lines in pull request are covered.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1364      +/-   ##
==========================================
- Coverage   93.96%   93.95%   -0.02%     
==========================================
  Files          52       52              
  Lines        7291     7291              
==========================================
- Hits         6851     6850       -1     
- Misses        440      441       +1     

Impacted Files	Coverage Δ
joblib/func_inspect.py	92.34% <100.00%> (ø)
joblib/test/test_store_backends.py	91.42% <0.00%> (-5.72%)	⬇️
joblib/pool.py	87.80% <0.00%> (-0.82%)	⬇️
joblib/parallel.py	96.03% <0.00%> (-0.53%)	⬇️
joblib/_parallel_backends.py	93.01% <0.00%> (ø)
joblib/test/test_parallel.py	97.05% <0.00%> (ø)
joblib/memory.py	95.51% <0.00%> (+0.26%)	⬆️
joblib/test/testutils.py	100.00% <0.00%> (+50.00%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[VaelK]

The doc build failed because of an URL not reached:
urllib.error.URLError: <urlopen error [Errno 104] Connection reset by peer>
It is not related to my work. Rerunning the workflow should fix this.

[VaelK]

Ok i think that some generated code in readthedocs tries to load a csv from an extrernal source (url) but fails since the server send a 104 error... I have no idea how to track this since I don't know where the code is generated. If one has any clue about that...

[ogrisel]

Could you please try to craft a minimal non-regression test for this fix?

[ogrisel]

Let's ignore the readthedocs failure for now. If it still happens let's fix it in an dedicated PR and keep this PR focused on the original issue.

[VaelK]

Ok, fine for me.
Thanks!

[jjerphan]

To complete @ogrisel's remark in #1364 (comment) here is scikit-learn's section on crafting minimal reproducer.

A non-regression test can then be added in this PR using this reproducer.

[VaelK]

Here is a minimal reproducer using shell command on linux with anaconda for setting the virtual environnement:

conda create --name test_joblib_issue_1364 python=3.9
conda activate test_joblib_issue_1364
pip install pytest==7.2.2
pip install gitpython==3.1.31
mkdir joblib.issue.1364
cd joblib.issue.1364/
git clone https://github.com/joblib/joblib.git
cd joblib
pytest joblib/test/test_memory.py
cd ../..
rm -rf joblib.issue.1364
conda remove --name test_joblib_issue_1364 --all

Expected result:
============================= 52 passed, 2 skipped, 4 warnings in 4.10s ====================================

Actual result:
====================================== short test summary info =========================================
FAILED joblib/test/test_memory.py::test_parallel_call_cached_function_defined_in_jupyter[True] - AssertionError: assert 'ipython-input' in 'main--home-baba-projets_git-joblib'
FAILED joblib/test/test_memory.py::test_parallel_call_cached_function_defined_in_jupyter[False] - AssertionError: assert 'ipython-input' in 'main--home-baba-projets_git-joblib'
=========================== 2 failed, 50 passed, 2 skipped, 4 warnings in 4.48s ================================

I am crafting a non regression test now.