feat: Add TNG to the list of readers

[jandom]

Maybe fixes https://github.com/MDAnalysis/UserGuide/issues/243 - could it be so simple?

[hmacdope]

You will also need to add a a page to reference the :ref: TNG against, see here: https://github.com/MDAnalysis/UserGuide/tree/6b82e9460e7b66ba742a17f2569842317b4cdf3b/doc/source/formats/reference

[lilyminium]

I think most .txt files are generated from scripts, this one being https://github.com/MDAnalysis/UserGuide/blob/develop/doc/source/scripts/gen_format_overview_classes.py .

There's also https://github.com/MDAnalysis/UserGuide/blob/develop/doc/source/scripts/core.py that needs to get updated.

[lilyminium]

Turns out scripts weren't running -- I opened #278 to fix this, but just warning that you might not see changes until that's in. If that's a blocker, one solution is to test locally by manually running the script itself.

[jandom]

Oh wow thanks for catching this @lilyminium, will wait – don't think this is pressing

[jandom]

@lilyminium not sure it helps but a single line change in core.py made this error go away and re-created a bunch of files. I think this (potentially?) goes hand-in-hand with your PR

Traceback (most recent call last):
  File "/Users/jandom/workspace/MDAnalysis/UserGuide/doc/source/scripts/gen_format_overview_classes.py", line 112, in <module>
    ov = FormatOverview()
  File "/Users/jandom/workspace/MDAnalysis/UserGuide/doc/source/scripts/base.py", line 36, in __init__
    self.get_lines(*args, **kwargs)
  File "/Users/jandom/workspace/MDAnalysis/UserGuide/doc/source/scripts/base.py", line 54, in get_lines
    lines.append(self.get_line(*items))
  File "/Users/jandom/workspace/MDAnalysis/UserGuide/doc/source/scripts/base.py", line 66, in get_line
    line.append(self._run_method(h, *args))
  File "/Users/jandom/workspace/MDAnalysis/UserGuide/doc/source/scripts/base.py", line 42, in _run_method
    val = meth(*args, **kwargs)
  File "/Users/jandom/workspace/MDAnalysis/UserGuide/doc/source/scripts/gen_format_overview_classes.py", line 55, in _description
    return DESCRIPTIONS[self.keys[-1]]
KeyError: 'TNG'

[lilyminium]

Yes, that's the key change that this PR needs to implement to fix the lack of TNG (the changes to the txt table should just get overwritten). It would add the TNG entry to the format overview table. However, if we have more to say about TNG, adding TNG.rst is necessary for the link to be functional. I tested the change you propose in this commit commit (https://github.com/MDAnalysis/UserGuide/pull/278/commits/687b9a8c2beaf198c47d230988364a4eda8b7000) in #278 -- it does indeed allow the docs building test to pass. (the examples test failed for unrelated reasons)

[jandom]

Okay this makes partial sense: there was some silent failure to generate some flies, because TNG was missing. I don't think I should be committing format_overview.txt or coordinate_readers.txt as these seem to be auto-generated?

Does anybody have a copy (the text we want to put into) the tng.rst file? Or should I just chat GPT their paper abstract into a 2-3 sentence summary?

[lilyminium]

@hmacdope do you have anything you'd like to say? These pages often document quirks, limitations, and assumptions made by the MDAnalysis parsers.

Or should I just chat GPT their paper abstract into a 2-3 sentence summary?

Otherwise this seems like a great alternative 😂

[jandom]

@hmacdope PLAL, i'm not sure what makes most sense to put in the TNG.rst, over what's already there in the MDA docs

[jandom]

@hmacdope bump, sorry to bug, PTAL, just because this is now blocking two PRs #280 and #271

[hmacdope]

@lilyminium given your comment in #280 I will let you decide when merging this is appropriate. My review cover the content. :)

[jandom]

Many thanks @hmacdope !

[jandom]

@lilyminium I've merged this one myself – this may have been a faux pas, because I don't know what's the merging custom. I found it quite surprising that others merged my PRs, but maybe that's totally normal here and I'm not supposed to merge (instead a maintainer is only allowed)

[lilyminium]

No worries at all @jandom, in the core library the maintainers often merge and I carried the habit over. Thanks for the fix, good to see CI green again!