Oliver Beckstein

If there's no decorator in there anymore then please update the PR comment and also reference issue #3651.

We have to update this everywhere at some point. It's just such an intrusive change in the history that affects all open PRs that we basically never want to do...

Can you add to the issue report what the correct answer is?

@jaclark5 if you've already identified a fix, can I encourage you to create a PR? We have some documention in the [User Guide on Contributing to the Main Code Base](https://userguide.mdanalysis.org/stable/contributing_code.html).

As this seems to be a broader problem with a common solution you can do one PR.

I agree, we have more experience now with making deps optional.

Hi @JoaoRodrigues , new formats are a great addition. Can you just create a PR and then we can comment directly on the code, run tests, etc?

Pure python implementation: https://github.com/Electrostatics/mmcif_pdbx

Also, [chemfiles reads mmCIF](http://chemfiles.org/chemfiles/latest/formats.html) and can be used inside MDAnalysis, see [Reading trajectories with chemfiles](https://docs.mdanalysis.org/2.0.0-dev0/documentation_pages/coordinates/chemfiles.html).

Note on trying to use chemfiles: With the 2.0.0b, ```python u = mda.Universe("1ake.cif.gz", topology_format="chemfiles") ``` fails with 'CHEMFILES' isn't a valid topology format, nor a coordinate format I can't just...