Skip to content

dependency bump: metatensor 0.9, metatomic 0.1.14, featomic 0.7.4#1144

Merged
pfebrer merged 27 commits into
metatensor:mainfrom
sofiia-chorna:bump-deps-metatensor-0.9
May 21, 2026
Merged

dependency bump: metatensor 0.9, metatomic 0.1.14, featomic 0.7.4#1144
pfebrer merged 27 commits into
metatensor:mainfrom
sofiia-chorna:bump-deps-metatensor-0.9

Conversation

@sofiia-chorna
Copy link
Copy Markdown
Collaborator

@sofiia-chorna sofiia-chorna commented May 15, 2026

reopen from #1128

dump to:

metatensor-learn >=0.5.0,<0.6
metatensor-operations >=0.5.0,<0.6
metatensor-torch >=0.9.0,<0.10
metatomic-torch >=0.1.14,<0.2
metatomic-ase >=0.1.1,<0.2
featomic-torch >=0.7.4,<0.8

Please go the the Preview tab and select the appropriate PR template:


📚 Documentation preview 📚: https://metatrain--1144.org.readthedocs.build/en/1144/

@Luthaf
Copy link
Copy Markdown
Member

Luthaf commented May 18, 2026

Featomic releases are now live on PyPI!

@pfebrer
Copy link
Copy Markdown
Contributor

pfebrer commented May 18, 2026

The errors are super strange, they come from importing torch 😅

@pfebrer
Copy link
Copy Markdown
Contributor

pfebrer commented May 18, 2026

And in soap-bpnn we have problems, can they be because of torch-spex?

@pfebrer
Copy link
Copy Markdown
Contributor

pfebrer commented May 18, 2026

Well, we have advanced to the next bug 😄

@Luthaf
Copy link
Copy Markdown
Member

Luthaf commented May 18, 2026

For the doc error I'm not sure where it comes from, but I'd fix the deprecation warnings and then check again.

For the windows test build, it looks like the wheel for sphericart-torch fails to build on Windows, and this part of the code has been touched recently. I'd try to run with sphericart 1.0.8 to see if this helps, and otherwise we'll need to enable more verbose logs on CI. Also ping @frostedoyster in case you have an idea about what's going on here!

pfebrer added 2 commits May 18, 2026 18:35
Removed version constraint for 'featomic' package.
@pfebrer
Copy link
Copy Markdown
Contributor

pfebrer commented May 19, 2026

@sofiia-chorna could you try to build the documentation locally and see if you understand what's happening?

@sofiia-chorna
Copy link
Copy Markdown
Collaborator Author

Traceback
=========

    Traceback (most recent call last):
      File "/Users/sofiia/cosmo/metatrain-fork/.tox/docs/lib/python3.11/site-packages/sphinx/cmd/build.py", line 414, in build_main
        app = Sphinx(
              ^^^^^^^
      File "/Users/sofiia/cosmo/metatrain-fork/.tox/docs/lib/python3.11/site-packages/sphinx/application.py", line 305, in __init__
        self.config.setup(self)
      File "/Users/sofiia/cosmo/metatrain-fork/docs/src/conf.py", line 92, in setup
        setup_architectures_docs()
      File "/Users/sofiia/cosmo/metatrain-fork/docs/src/architectures/generate.py", line 157, in setup_architectures_docs
        preload_documentation_module(architecture_name)
      File "/Users/sofiia/cosmo/metatrain-fork/.tox/docs/lib/python3.11/site-packages/metatrain/utils/architectures.py", line 233, in preload_documentation_module
        spec.loader.exec_module(documentation)
      File "<frozen importlib._bootstrap_external>", line 940, in exec_module
      File "<frozen importlib._bootstrap>", line 241, in _call_with_frames_removed
      File "/Users/sofiia/cosmo/metatrain-fork/.tox/docs/lib/python3.11/site-packages/metatrain/pet/documentation.py", line 73, in <module>
        from metatrain.pet.modules.finetuning import FinetuneHypers, NoFinetuneHypers
      File "/Users/sofiia/cosmo/metatrain-fork/.tox/docs/lib/python3.11/site-packages/metatrain/pet/__init__.py", line 1, in <module>
        from .model import PET
      File "/Users/sofiia/cosmo/metatrain-fork/.tox/docs/lib/python3.11/site-packages/metatrain/pet/model.py", line 43, in <module>
        class PET(ModelInterface[ModelHypers]):
      File "/Users/sofiia/cosmo/metatrain-fork/.tox/docs/lib/python3.11/site-packages/metatrain/pet/model.py", line 58, in PET
        __default_metadata__ = ModelMetadata(
                               ^^^^^^^^^^^^^^
      File "/Users/sofiia/cosmo/metatrain-fork/.tox/docs/lib/python3.11/site-packages/metatomic/torch/documentation.py", line 491, in __init__
        raise THIS_CODE_SHOULD_NOT_RUN
    RuntimeError: This is the documentation-only version of this class, it should not be used at runtime. If you see this error while running examples, check above for the actual error message.

@Luthaf
Copy link
Copy Markdown
Member

Luthaf commented May 19, 2026

Is there anything else above?

@sofiia-chorna
Copy link
Copy Markdown
Collaborator Author

sofiia-chorna commented May 19, 2026

examples seem to be generated cleanly but it crashes in setup_architectures_docs(). let me attach logs:
tox-docs.log

@sofiia-chorna
Copy link
Copy Markdown
Collaborator Author

did not manage to setup it on CI (torch cpu index) so far! on CI only tox -e tests fails due to setup, locally it passes:
tox-tests-override-rerun.log

@Luthaf Luthaf force-pushed the bump-deps-metatensor-0.9 branch from 449c925 to bc5f33d Compare May 20, 2026 08:02
@Luthaf Luthaf force-pushed the bump-deps-metatensor-0.9 branch from bc5f33d to 8394fd3 Compare May 20, 2026 08:14
@sofiia-chorna
Copy link
Copy Markdown
Collaborator Author

sofiia-chorna commented May 20, 2026

i think it remains to have the release of metatomic v0.1.14 publised metatensor/metatomic#231 and updated here and we are good?

readthedocs fails due to the same issue as CI (fixed with temporal use of torch cpu index)


    Traceback (most recent call last):
      File "/home/docs/checkouts/readthedocs.org/user_builds/metatrain/envs/1144/lib/python3.13/site-packages/sphinx/cmd/build.py", line 414, in build_main
        app = Sphinx(
            srcdir=args.sourcedir,
        ...<14 lines>...
            exception_on_warning=args.exception_on_warning,
        )
      File "/home/docs/checkouts/readthedocs.org/user_builds/metatrain/envs/1144/lib/python3.13/site-packages/sphinx/application.py", line 305, in __init__
        self.config.setup(self)
        ~~~~~~~~~~~~~~~~~^^^^^^
      File "/home/docs/checkouts/readthedocs.org/user_builds/metatrain/checkouts/1144/docs/src/conf.py", line 92, in setup
        setup_architectures_docs()
        ~~~~~~~~~~~~~~~~~~~~~~~~^^
      File "/home/docs/checkouts/readthedocs.org/user_builds/metatrain/checkouts/1144/docs/src/architectures/generate.py", line 157, in setup_architectures_docs
        preload_documentation_module(architecture_name)
        ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^
      File "/home/docs/checkouts/readthedocs.org/user_builds/metatrain/envs/1144/lib/python3.13/site-packages/metatrain/utils/architectures.py", line 233, in preload_documentation_module
        spec.loader.exec_module(documentation)
        ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^
      File "<frozen importlib._bootstrap_external>", line 1027, in exec_module
      File "<frozen importlib._bootstrap>", line 488, in _call_with_frames_removed
      File "/home/docs/checkouts/readthedocs.org/user_builds/metatrain/envs/1144/lib/python3.13/site-packages/metatrain/pet/documentation.py", line 73, in <module>
        from metatrain.pet.modules.finetuning import FinetuneHypers, NoFinetuneHypers
      File "/home/docs/checkouts/readthedocs.org/user_builds/metatrain/envs/1144/lib/python3.13/site-packages/metatrain/pet/__init__.py", line 1, in <module>
        from .model import PET
      File "/home/docs/checkouts/readthedocs.org/user_builds/metatrain/envs/1144/lib/python3.13/site-packages/metatrain/pet/model.py", line 43, in <module>
        class PET(ModelInterface[ModelHypers]):
        ...<1414 lines>...
                return checkpoint
      File "/home/docs/checkouts/readthedocs.org/user_builds/metatrain/envs/1144/lib/python3.13/site-packages/metatrain/pet/model.py", line 58, in PET
        __default_metadata__ = ModelMetadata(
            references={"architecture": ["https://arxiv.org/abs/2305.19302v3"]}
        )
      File "/home/docs/checkouts/readthedocs.org/user_builds/metatrain/envs/1144/lib/python3.13/site-packages/metatomic/torch/documentation.py", line 491, in __init__
        raise THIS_CODE_SHOULD_NOT_RUN
    RuntimeError: This is the documentation-only version of this class, it should not be used at runtime. If you see this error while running examples, check above for the actual error message.

@sofiia-chorna sofiia-chorna changed the title dependency bump: metatensor 0.9 + metatomic 0.1.12 dependency bump: metatensor 0.9, metatomic 0.1.14, featomic 0.7.4 May 20, 2026
@sofiia-chorna sofiia-chorna marked this pull request as ready for review May 20, 2026 15:07
@Luthaf
Copy link
Copy Markdown
Member

Luthaf commented May 20, 2026

cscs-ci run

Copy link
Copy Markdown
Contributor

@pfebrer pfebrer left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ok, a couple things and we are good 👍

Comment thread pyproject.toml Outdated
Comment thread pyproject.toml
Comment on lines +202 to +214
"ignore:`compute_requested_neighbors_from_options` is deprecated and will be removed in a future version:UserWarning",
# metatomic 0.1.12 deprecations
"ignore:the 'features' output name is deprecated:UserWarning",
"ignore:the 'features' quantity is deprecated:UserWarning",
"ignore:the 'features' quantity is deprecated:DeprecationWarning",
"ignore:the 'non_conservative_forces' output name is deprecated:UserWarning",
"ignore:the 'momenta' quantity is deprecated:DeprecationWarning",
"ignore:the 'positions' quantity is deprecated:DeprecationWarning",
"ignore:the 'masses' quantity is deprecated:DeprecationWarning",
"ignore:the 'velocities' quantity is deprecated:DeprecationWarning",
"ignore:the 'charges' quantity is deprecated:DeprecationWarning",
"ignore:`per_atom` is deprecated:DeprecationWarning",
"ignore:ModelOutput.quantity is deprecated:DeprecationWarning"
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We have to get rid of all these warnings ASAP after this PR 😅

Copy link
Copy Markdown
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ok i will do it in the follow-up PR

Comment thread src/metatrain/gap/model.py Outdated
@sofiia-chorna sofiia-chorna force-pushed the bump-deps-metatensor-0.9 branch from b16bfae to 86ed229 Compare May 20, 2026 16:08
@sofiia-chorna
Copy link
Copy Markdown
Collaborator Author

cscs-ci run

@Luthaf
Copy link
Copy Markdown
Member

Luthaf commented May 21, 2026

cscs-ci run

@pfebrer pfebrer merged commit 331f676 into metatensor:main May 21, 2026
21 checks passed
@pfebrer
Copy link
Copy Markdown
Contributor

pfebrer commented May 21, 2026

Thank youuuu

JonathanSchmidt1 added a commit to JonathanSchmidt1/metatrain that referenced this pull request May 21, 2026
…lse) merge

The reference fixture's storage changed slightly when the scaler switched
from the manual TensorBlock rebuild to block.copy(deep=False) (from PR
metatensor#1144). Regenerated via test_checkpoint_did_not_change.
JonathanSchmidt1 added a commit to JonathanSchmidt1/metatrain that referenced this pull request May 21, 2026
The merge with PR metatensor#1144 replaced our regex catch-alls for torch.jit
deprecations and the plural-name family ('features', 'positions',
'momenta', etc.) with PR metatensor#1144's narrower per-warning entries. macOS CI
was failing 61 tests on a metatomic 0.1.14 UserWarning about
Model.requested_inputs(use_new_names=False) that no remaining filter
caught.

Restore the regex catch-alls, add an explicit ignore for the new
requested_inputs deprecation, and fix the one offending call site
(cli/eval.py) to pass use_new_names=True so we no longer emit the
warning from our own code.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants