First steps towards the trajectory endpoint (Johan's legacy)

[d-beltran]

Hi everyone,

I tried to rescue as much code as possible from an old branch in the Johan personal repository of optimade python tools.
I removed all the compatibility layer which was specific for our database and fixed most compatibility issues with the current optimade version but there is one last failing test.

You may reproduce the problem running pytest tests/server/test_server_validation.py
The error is giving me no clue of what I should do to fix it and I have spend already much time trying to.
I think it is some issue related to queries aiming for 'frame_serialization_format' and 'nvalues', which are two trajectory attributes declared under the 'AvailablePropertySubfields' class. These are declared apart and they are not in the current trajectories specification, so we could simply delete them, but I am afraid there is some error behind this and it will appear back when we declare new property subfields, so it should be understood and fixed before starting to implement the new specification.

This pull request is being done in collaboration with @ml-evs and @ndaelman-hu.

[JPBergsma]

Hi Daniel,

Good to see that you are continuing my work,
If I remember correctly, the specification for trajectories and such was still changed drastically, at the last optimade meeting where I was present(I think June 2023). So unless those changes were reverted, the branch that I used for your project does not follow the latest version of the Optimade specification. There are multiple branches on my gitrepo that implement those features.
I also had a trajectory branch that I based your version of the optimade python tools on. So it should not be necessary to remove the features specific for your database.
I will move back to Switzerland in two weeks, so I am a bit busy now. But I'll try to look tonight when I am home to see which branches are the branches you should continue from.

[JPBergsma]

The following branches are the ones that are relevant for specification as discussed at the OPTIMADE meeting in June 2023:
https://github.com/JPBergsma/optimade-python-tools/tree/JPBergsma/Partial_data
https://github.com/JPBergsma/optimade-python-tools/tree/JPBergsma/trajectory_endpoint
https://github.com/JPBergsma/optimade-python-tools/tree/JPBergsma/add_metadata

The original branch on which the bioexcel is based is:
https://github.com/JPBergsma/optimade-python-tools/tree/optimade_python_tools_trajectory_0.1
Although it is based on an earlier version of the Optimade specification.

Let me know if you have any more questions.

[ml-evs]

Thanks for this @d-beltran (and great to hear from you @JPBergsma!) -- I'll try to keep an eye on this PR, but please ping me on here or over email if there is anything you need help with.

[d-beltran]

Hey Johan :)

The branch I started from was another one:
https://github.com/JPBergsma/optimade-python-tools/tree/JPBergsma_BioExcel

Thanks for the clues!