add grimgrains.com

[shantih19]

Add scraper for https://grimgrains.com

[jayaddison]

Looks great to me, thanks @shantih19

One annoying detail that isn't specific to this pull request, but a problem with the codebase at the moment: we seem to have a mix of types in the instructions field results.

In the past it had always returned a newline-separated (single) string value, but nowadays we have a few scrapers that return a list of strings.

I should raise a separate issue thread about that and figure out what to do This has been addressed by pull request #852. I'll merge this soon either way - this is mostly a heads-up that that field is a bit ambiguous type-wise at the moment.

[shantih19]

I see, if you prefer, I can change the format to the newline-separated string

[jayaddison]

I see, if you prefer, I can change the format to the newline-separated string

Yep, let's do that. I think the first step will be to make everything consistent, so that'll mean one less method to remember to update.

It's slightly annoying that mypy doesn't seem to detect this automatically. I guess if there's no type hint on the subclass methods, it treats them as untyped (the abstract class does define a return type...).

[jayaddison]

Relevant mypy bug seems to be: python/mypy#3903

[shantih19]

I'll add the None return when the string is empty (like the edits in #852)

[jayaddison]

I see, if you prefer, I can change the format to the newline-separated string

Yep, let's do that. I think the first step will be to make everything consistent, so that'll mean one less method to remember to update.

This consistency mini-project should be complete: all of the scrapers in the codebase now return values of string type (str) from the instructions method.

[jayaddison]

I think we should remove the description method here - apart from that, this looks good to go 👍

[jayaddison]

Note-to-future: this url attribute is set because we return a fully-qualified image URL, and when the tests run (on HTML, outside of a URL context) the base URL is test.example.com, which looks wrong.

[jayaddison]

Thanks @shantih19 - this looks great, and ready for merge. Thanks for your patience with the review and main-branch code edits during this!