WIP: Auto-Import User-Defined Cosmology Realizations - #11517

👋 Thank you for your draft pull request! Do you know that you can use [ci skip] or [skip ci] in your commit messages to skip running continuous integration tests until you are ready?

Uh oh! I forgot PyYAML is not a run-time dependency... how can an ECSV file be read?

Might have to re-think this implementation. Sigh. JSON files, I guess ?

@taldcroft, any suggestions how to avoid JSON and keep ECSV?

@nstarman - unfortunately I can't think of a way to use ECSV for this case. It really does require YAML, so the choice is either don't use ECSV or make YAML a requirement for cosmology (and hence astropy).

BTW, if you are wondering about why we used YAML not JSON for ECSV.. I often think about this myself:

• A key showstopper is that JSON does not allow numeric values for a dict key. So it cannot represent arbitrary pure-Python data structures.
• Much better system for extensibility.
• More human-readable.

I spent a little time thinking about ways of reading key bits of the YAML header "by hand" without YAML, but at the end I don't think this will be worth it. Sigh.

BTW, 👍 on the idea, even if it ends up not working out!

BTW, 👍 on the idea, even if it ends up not working out!

Thanks!

unfortunately I can't think of a way to use ECSV for this case. It really does require YAML, so the choice is either don't use ECSV or make YAML a requirement for cosmology (and hence astropy).

Well, I don't want to add any requirements... So I guess I'll use JSON for astropy's internal storage 😢 .
I think I'll add methods Cosmology.read/write for JSON I/O.

For user-defined cosmologies, I think the ECSV is more logical and the code I have here will work well. Perhaps I'll just transition this PR into one for user-defined cosmologies.

Thanks @taldcroft for the quick reply.

Hello @nstarman 👋! It looks like you've made some changes in your pull request, so I've checked the code again for style.

• In the file astropy/cosmology/parameters.py:

Line 273:1: W293 blank line contains whitespace

• In the file astropy/cosmology/tests/test_parameters.py:

Line 9:1: E302 expected 2 blank lines, found 1

Comment last updated at 2021-04-28 20:07:24 UTC

I guess I'll use JSON

What about FITS? 😆

I'm also planning on Linear A as soon as it's deciphered.

@nstarman At some point I'd like to have a discussion to better wrap my head around your cosmology plans. I like where this is headed though and I think you have the right idea. +1 to using YAML to describe cosomologies, and if we need to add pyyaml as a dependency for astropy I would not be opposed.

@embray Adding PyYAML would be awesome. @taldcroft and I had some discussion in a different PR about this... and it would be very nice to have a clear way of serializing quantities.
I would be more than happy to chat about my cosmology plans... perhaps that would help me understand them as well! 😆. I've defined some broad goals, but as you can see from the various WIP PRs, I haven't yet coordinated a specific line of attack. Would this be best done on Slack? There's the languishing cosmology channel.

Worth doing, but this was way outdated and the wrong approach.