Run regression tests on nightlies (PyPI CPU, GPU. & Conda CPU, GPU)

[agunapal]

Description

This PR adds the following

• Workflow to test TorchServe CPU pypi & conda binaries for Python 3.8, 3.9, 3.10
• Workflow to test TorchServe GPU pypi & conda binaries for Python 3.8, 3.9, 3.10

Thanks to @udaij12 for helping me get conda working with the GPU runner
Fixes #(issue)

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• New feature (non-breaking change which adds functionality)
• This change requires a documentation update

Feature/Issue validation/testing

CPU run : https://github.com/pytorch/serve/actions/runs/6132696896/job/16643849251?pr=2562
GPU run: https://github.com/pytorch/serve/actions/runs/6132974703/job/16645443959?pr=2562

CPU Conda

Mac Py 3.10 fails because we didn't have any binaries before.
6_regression-cpu-conda-nightly (macOS-latest, 3.10).txt
4_regression-cpu-conda-nightly (macOS-latest, 3.8).txt
1_regression-cpu-conda-nightly.txt
2_regression-cpu-conda-nightly (ubuntu-20.04, 3.9).txt
1_regression-cpu-conda-nightly (ubuntu-20.04, 3.8).txt
3_regression-cpu-conda-nightly (ubuntu-20.04, 3.10).txt
5_regression-cpu-conda-nightly (macOS-latest, 3.9).txt

CPU PyPI

4_regression-cpu-pypi-nightly (macOS-latest, 3.8).txt
3_regression-cpu-pypi-nightly (ubuntu-20.04, 3.10).txt
6_regression-cpu-pypi-nightly (macOS-latest, 3.10).txt
2_regression-cpu-pypi-nightly (ubuntu-20.04, 3.9).txt
1_regression-cpu-pypi-nightly (ubuntu-20.04, 3.8).txt
5_regression-cpu-pypi-nightly (macOS-latest, 3.9).txt

GPU PyPI

1_regression-gpu-pypi-nightly (3.8).txt
2_regression-gpu-pypi-nightly (3.9).txt
1_regression-gpu-pypi-nightly.txt
3_regression-gpu-pypi-nightly (3.10).txt

Checklist:

• Did you have fun?
• Have you added tests that prove your fix is effective or that this feature works?
• Has code been commented, particularly in hard-to-understand areas?
• Have you made corresponding changes to the documentation?

[codecov]

Codecov Report

Merging #2562 (627f2d7) into master (5e30a7d) will decrease coverage by 0.08%.
The diff coverage is n/a.

❗ Current head 627f2d7 differs from pull request most recent head 20ce65e. Consider uploading reports for the commit 20ce65e to get more accurate results

@@            Coverage Diff             @@
##           master    #2562      +/-   ##
==========================================
- Coverage   70.87%   70.79%   -0.08%     
==========================================
  Files          83       83              
  Lines        3839     3839              
  Branches       58       58              
==========================================
- Hits         2721     2718       -3     
- Misses       1114     1117       +3     
  Partials        4        4              

see 1 file with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[msaroufim]

could you plz elaborate on the needed changes for the newman tests?

[agunapal]

I am not sure i follow. There is no change to the tests. I am only changing how we install torchserve

[agunapal]

Git diff makes it seem like I have changed the tests

[msaroufim]

Hrmm the new code owner policy isnt' great considering i can't merge this unless Matthias checks the tests

[mreso]

Just a few nits, other than that LGTM

[mreso]

nit: Can we not derive thats its binary if pypi or conda argument is given?

[mreso]

Is there any value in running the command with try_and_handle here? The only exception it handles, its handling by throwing it again, and we do not use dry_run here, so why not call the command directly?