refactor!: config parsing

[tristan-f-r]

Closes #299. I did want to keep this separate from #228 so that way those details didn't get lost, but this does the proper work of using structural parsing on the config, throwing useful errors on missing keys, wrongly-written keys, default paramaters, and with proper input validation and casting.

This doesn't clean up the parsing done in #228, but it lays the ground for cleaner configuration parsing.

I've also made some issues for suggestions for the configuration format #297 #298 #299 #309.
For users, this affects error messages: Now, we get good error messages for missing keys and too many keys, or improperly formatted values.

Note

The breaking change indicator in the PR header (!) is present because this will cause the config to raise an error if there are too many keys present inside any one of the objects (except for the runs... object). This also removes reconstruction_settings -> run.

[tristan-f-r]

It turns out that we had a good chunk of unspecified defaults in the config that are now being heavily caught.

[annaritz]

Suggestion: rename raw_config to config_specs. Keep config.py as named for now, but a later refactor might be in order.

@agitter if you can take a look at this and give some comments, that would be great!

[read-the-docs-community]

Documentation build overview

📚 spras | 🛠️ build #29005075 (e82eddf) | 🔍 preview

Files changed

Comparing with latest (32a0424...e82eddf)

Show files (3) | 3 modified | 0 added | 0 deleted

File	Status
genindex.html	📝 modified
fordevs/modules.html	📝 modified
fordevs/spras.html	📝 modified

[agitter]

The additional comments are helpful for me. I don't have any remaining major concerns. I am not sure how to test this beyond the test cases we have.

Neha, do you want to do more review before we merge?

[ntalluri]

I’ll do another review before merging.

[agitter]

I’ll do another review before merging.

We could also discuss this and the follow up PRs at our Monday meeting if necessary

[ntalluri]

@agitter Tristan addressed all of my remaining questions. This is ready to merge, but we should hold off until we look over the other PRs today in case we decide to wait on merging.

[tristan-f-r]

(Is that to avoid merge conflicts?)

[ntalluri]

I have a feeling we’ll end up merging it later today, but it would be helpful to first have a conversation about all the PRs that depend on this one. There may be a reason not to merge it yet, though nothing comes to mind right now.

[tristan-f-r]

I think that's fine to merge then? I want to split off an independent chunk of #329 which depends on this PR that should lighten the diff on #329 a little (about 13 lines of actual code with a 1k line artifact), and begin writing some config documentation using this PR.

[tristan-f-r]

Actually, I'll merge upstream first. (done, pre-commit passed 👍)

[tristan-f-r]

I want to split off an independent chunk of #329 which depends on this PR that should lighten the diff on #329 a little (about 13 lines of actual code with a 1k line artifact),

#358.

[agitter]

it would be helpful to first have a conversation about all the PRs that depend on this one

That was to understand the dependent changes and how much further restructuring was coming and what new maintenance burden, if any, would be introduced.