TorchANI 2.2.1 - Githubissues

yueyericardo commented 2 years ago

Checklist

[x] Used a personal fork of the feedstock to propose changes
[x] Bumped the build number (if the version is unchanged)
[x] Reset the build number to 0 (if the version changed)
[x] Re-rendered with the latest conda-smithy (Use the phrase code>@<space/conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
[x] Ensured the license file is being packaged.

Fix #1, torchani 2.2.1 with cuda_extension support for cuaev.

conda-forge-linter commented 2 years ago

Hi! This is the friendly automated conda-forge-linting service.

I wanted to let you know that I linted all conda-recipes in your PR (recipe) and found some lint.

Here's what I've got...

For recipe:

The top level meta keys are in an unexpected order. Expecting ['package', 'source', 'build', 'requirements', 'test', 'about', 'extra'].

For recipe:

Jinja2 variable references are suggested to take a {{<one space><variable name><one space>}} form. See lines [48].

yueyericardo commented 2 years ago

@conda-forge-admin, please rerender

conda-forge-linter commented 2 years ago

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

I do have some suggestions for making it better though...

For recipe:

Jinja2 variable references are suggested to take a {{<one space><variable name><one space>}} form. See lines [15].

yueyericardo commented 2 years ago

@conda-forge-admin, please rerender

conda-forge-linter commented 2 years ago

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

yueyericardo commented 2 years ago

@conda-forge-admin, please rerender

yueyericardo commented 2 years ago

Hi @hadim, this PR is ready to review.

yueyericardo commented 2 years ago

Re-trigger test

hmaarrfk commented 2 years ago

FYI seems like we screwed up our run export on pytorch.

https://github.com/conda-forge/pytorch-cpu-feedstock/blob/master/recipe/meta.yaml

We added it to the parent recipe, but not the outputs. sorry for that. You should definitely pin yourself in the requires.

hadim commented 2 years ago

Assuming the latest commit (pinning pytorch in run) is correct, LGTM for me here. @hmaarrfk can you make a last quick review in case I missed anything?

Thank you @yueyericardo for the work here!

I'll merge once CI is green and @hmaarrfk confirm everything is fine.

conda-forge-linter commented 2 years ago

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

I do have some suggestions for making it better though...

For recipe:

Jinja2 variable references are suggested to take a {{<one space><variable name><one space>}} form. See lines [22].

conda-forge-linter commented 2 years ago

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

hmaarrfk commented 2 years ago

I think this discussion is better had on the conda-forge-pinnings repo. you could use this as an example of the perils of not dropping 10.2 or not having a global pinning for pytorch (both of which are in PRs now ready for review)

hadim commented 2 years ago

Let's wait for https://github.com/conda-forge/staged-recipes/pull/16698 to be merged and then I think we can merge here (after switching from pkbar pip to the conda forge one).

hadim commented 2 years ago

@yueyericardo pkbar is now available on conda forge. Can you remove the pip install step and move pkbar into the conda deps?

yueyericardo commented 2 years ago

done and thanks for packaging pkbar to conda-forge!

hadim commented 2 years ago

Looks like certain builds fails because of a packaging issues with pytorch/cuda. I don't think this is related to torchani:

File "/opt/conda/lib/python3.8/site-packages/conda_build/render.py", line 379, in _read_upstream_pin_files
    extra_run_specs = get_upstream_pins(m, actions, env)
  File "/opt/conda/lib/python3.8/site-packages/conda_build/render.py", line 365, in get_upstream_pins
    loc, dist = execute_download_actions(m, actions, env=env, package_subset=pkg)[pkg]
  File "/opt/conda/lib/python3.8/site-packages/conda_build/render.py", line 330, in execute_download_actions
    pfe.execute()
  File "/opt/conda/lib/python3.8/site-packages/conda/core/package_cache_data.py", line 655, in execute
    raise CondaMultiError(exceptions)
conda.CondaMultiError: Downloaded bytes did not match Content-Length
  url: https://conda.anaconda.org/conda-forge/linux-64/pytorch-1.9.1-cuda110py39h5cf7045_3.tar.bz2
  target_path: /opt/conda/pkgs/pytorch-1.9.1-cuda110py39h5cf7045_3.tar.bz2
  Content-Length: 564081090
  downloaded bytes: 184777172

yueyericardo commented 2 years ago

All tests passed now @hadim

hadim commented 2 years ago

Thanks to everyone who participated here! Especially @yueyericardo.

conda-forge / torchani-feedstock

TorchANI 2.2.1 #2