Test installation in CI

ds4dm / ecole

Extensible Combinatorial Optimization Learning Environments

https://www.ecole.ai

BSD 3-Clause "New" or "Revised" License

319 stars 68 forks source link

Test installation in CI #267

Closed AntoinePrv closed 2 years ago

AntoinePrv commented 2 years ago

Pull request checklist

[x] I have opened an issue to discuss the proposed changes: Partial #136
[x] I have modified/added tests to cover the new changes/features.
[x] ~I have modified/added the documentation to cover the new changes/features.~
[x] I have ran the tests, checks, and code formatters.

Proposed implementation

This PR test that examples in the examples folder are tested with the the current ecole version.

Notebooks are tested using papermill. Conda environment files are added.
CMake example tests the installation of libecole

AntoinePrv commented 2 years ago

@gasse I've done some minor adjustment to the configuring tutorial, it's mostly reordering the global parameters. I've re-executed the notebook, let me know if the outputs looks decent.

@dchetelat similarily, I've done some adjustments to the branching example. I've also managed to install ecole, pytorch and pytorch_geometric in a single environment, but that was not easy. 😅 As a result this is now running the latest pytorch_geometric. The notebook execution fails, but I think it is not a big issue (something with None), could you have a look? Because of the error, there is currently no outputs for the notebook (if you feel like committing one 🙏 ).

---> 18         self.constraint_features = torch.FloatTensor(constraint_features)
TypeError: new(): data must be a sequence (got NoneType)

dchetelat commented 2 years ago

@dchetelat similarily, I've done some adjustments to the branching example. I've also managed to install ecole, pytorch and pytorch_geometric in a single environment, but that was not easy. sweat_smile As a result this is now running the latest pytorch_geometric. The notebook execution fails, but I think it is not a big issue (something with None), could you have a look? Because of the error, there is currently no outputs for the notebook (if you feel like committing one pray ).
---> 18         self.constraint_features = torch.FloatTensor(constraint_features)
TypeError: new(): data must be a sequence (got NoneType)

Yeah I checked, pyg 2.0.2 introduced a bunch of breaking changes, it's fixed now.