Upgrade the required python version to python>=3.11

[westford14]

PR to update the python required version to be >= 3.10 -- this required unpinning the thop dependency as well as changing the ModelEnum.time_distributed extraction of the string to ModelEnum.time_distributed.value.

Closes #310
Closes #192

[netlify]

✅ Deploy Preview for silly-keller-664934 ready!

Name	Link
🔨 Latest commit	20083ec
🔍 Latest deploy log	https://app.netlify.com/sites/silly-keller-664934/deploys/66e67223865ed600087c6636
😎 Deploy Preview	https://deploy-preview-337--silly-keller-664934.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site configuration.

[westford14]

@ejm714 -- let me know if this is the right direction for this PR

[ejm714]

@westford14 this is the right direction. Let's see if we can set the floor at 3.11 instead and then use 3.11 and 3.12 in the tests. You'll need to update the version here for the tests for github actions:

zamba/.github/workflows/tests.yml

Line 44 in 498bca1

python-version: [3.8, 3.9]

Once that's changed, I'll approve and run the github workflows for this PR

[westford14]

@westford14 this is the right direction. Let's see if we can set the floor at 3.11 instead and then use 3.11 and 3.12 in the tests. You'll need to update the version here for the tests for github actions:

zamba/.github/workflows/tests.yml

Line 44 in 498bca1

python-version: [3.8, 3.9]

Once that's changed, I'll approve and run the github workflows for this PR

@ejm714 -- thanks for the heads up on the github actions, i've now pinned them to 3.11 / 3.12

[ejm714]

Couple small things for user clarity

[westford14]

thanks for the comments @ejm714 -- i've updated accordingly

[westford14]

@ejm714 -- sorry i fixed the matrix sequence in the github actions file

[ejm714]

Thanks. That got past the initial hurdle so we actually get to run the tests on this PR however they're failing due to some installs for the densepose tests. You should be able to run the tests locally with make test for faster debugging. There's some details here: https://zamba.drivendata.org/docs/stable/contribute/#running-the-zamba-test-suite

[westford14]

looking at the pytorch issues, only python 3.12 support is only available in pytorch >== https://github.com/pytorch/pytorch/releases/tag/v2.4.0 which breaks the densepose library dependency, i guess the best path forward here is to pin to python 3.11

@ejm714 i've updated the code accordingly

[westford14]

@ejm714 -- pinned to 3.11

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 87.3%. Comparing base (498bca1) to head (20083ec).
Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff            @@
##           master    #337     +/-   ##
========================================
- Coverage    87.4%   87.3%   -0.1%     
========================================
  Files          26      26             
  Lines        2191    2191             
========================================
- Hits         1915    1913      -2     
- Misses        276     278      +2     

Files with missing lines	Coverage Δ
zamba/models/config.py	97.5% <100.0%> (ø)

... and 1 file with indirect coverage changes

[westford14]

hey @ejm714 -- any changes I need to make to this PR? looks like all the tests would have passed (the macOS 3.11 was cancelled)

[westford14]

wow codecov is really ratelimiting hard here

[ejm714]

hey @ejm714 -- any changes I need to make to this PR? looks like all the tests would have passed (the macOS 3.11 was cancelled)

It's unclear. I want to make sure all tests are passing on all operating systems but as you noted, codecov is really ratelimiting and I haven't had a chance to look into why/how to fix that. If you want to speed merging along, then you could look into that. Otherwise, feel free to put this down until I request anything specifically

[westford14]

hey @ejm714 -- any changes I need to make to this PR? looks like all the tests would have passed (the macOS 3.11 was cancelled)

It's unclear. I want to make sure all tests are passing on all operating systems but as you noted, codecov is really ratelimiting and I haven't had a chance to look into why/how to fix that. If you want to speed merging along, then you could look into that. Otherwise, feel free to pull this down until I request anything specifically

yea i'll take a look at the ratelimit thing there -- though it does seem like the tests are passing

[westford14]

@ejm714 -- was doing some research into the codecov issue and it seems to be a pretty common issue for forked repos, but in this thread someone mentioned that upgrading to the v4 github action resolved it (codecov/feedback#358 (comment))

figured this was worth a shot before fully re-architecting this

[ejm714]

Looks like we can actually support python 3.12

I was looking into this more and we need to remove the ceiling on torch in the densepose installs. The minimum torch version is 2.2.0 which is why we were getting the failures with python 3.12 and the densepose install. With the latest torch, the densepose tests are passing locally with python 3.12.

Thanks for the codecov update -- that seems to have resolved the timeouts. Once you've made the changes to support 3.11 and 3.12, I'll run the tests again and we should be good to go

[westford14]

Looks like we can actually support python 3.12

I was looking into this more and we need to remove the ceiling on torch in the densepose installs. The minimum torch version is 2.2.0 which is why we were getting the failures with python 3.12 and the densepose install. With the latest torch, the densepose tests are passing locally with python 3.12.

Thanks for the codecov update -- that seems to have resolved the timeouts. Once you've made the changes to support 3.11 and 3.12, I'll run the tests again and we should be good to go

awesome -- i'll do that and let you know when it's ready to rerun the tests

[westford14]

@ejm714 -- just removed the cap on the densepose pytorch version, tests passed locally for me in both 3.11 and 3.12

[ejm714]

Looks good, thanks!