Support Python 3.13 #227
Support Python 3.13 #227
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #227 +/- ##
===========================================
Coverage 100.00% 100.00%
===========================================
Files 45 60 +15
Lines 2242 3970 +1728
===========================================
+ Hits 2242 3970 +1728 ☔ View full report in Codecov by Sentry. |
chezou
force-pushed
the
feature/py313
branch
4 times, most recently
from
1152e98 to
384dfcc
Compare
chezou
force-pushed
the
feature/py313
branch
from
6e6d193 to
40e3b71
Compare
chezou
force-pushed
the
feature/py313
branch
from
40e3b71 to
d37e414
Compare
chezou
force-pushed
the
feature/py313
branch
from
23bfe8f to
2d5ccf3
Compare
| base_model = TFIDFRecommender(K=5, num_threads=2) | ||
| # Recreate dataset to prevent same co-occurrence count between (11, 14) and (11, 15) | ||
| # which leads to different results in the test in Python 3.13 |
There was a problem hiding this comment.
You mean for the default dataset the results in python 3.8-3.12 and 3.13 are different?
Do you have any idea why is this happening?
Probably the problem is not with the python itself, but with some of the dependencies?
There was a problem hiding this comment.
Yes, the behavior of implicit ItemKNN changes due to the changes in the similarity matrix.
Considering the cosign similarity, co-occurrence between 11-14 and 11-15 are both 2 times, thus it causes flakiness.
There was a problem hiding this comment.
Let us take some time to examine this case carefully
There was a problem hiding this comment.
@feldlime Do you need more time?
I think the original dataset can be a bit flaky because the cosine similarities for both two are the same.
There was a problem hiding this comment.
Didn't have time to work on this, I'll take another 2-3 weeks
|
@feldlime Applied all the feedback, and CI passed. PTAL |
Co-authored-by: Emiliy Feldman <[email protected]>
|
For nmslib, I created Python 3.13 support PR at metabrainz/nmslib-metabrainz#5 |
|
Test failure is due to upload coverage failure https://github.com/MobileTeleSystems/RecTools/actions/runs/12814286663/job/35730351993
|
|
PyTorch will support Python 3.13 on January 29. |
Description
Support Python 3.13 excluding:
PyTorch will likely support Python 3.13 in the coming months.
nmslib-metabrainz only supports until Python 3.12.
Type of change
How Has This Been Tested?
Before submitting a PR, please check yourself against the following list. It would save us quite a lot of time.