Fix all i18n issues detected by new script

[rebecca-shoptaw]

Closes #9486.

Fix. Manually fixes all i18n syntax issues picked up by the new pre-commit script. Also escapes some non-i18n-able and/or accidentally detected lines.

Technical

• Involved adjusting the warn regex so that it detects $(' or $(" to avoid creating WARNs for Python code inserts
• I'm now seeing a SyntaxWarning when the new messages.pot is generated; would be nice to get this fixed before merge, but everything still runs as expected
• Had to manually escape some Python inserts that began with quotes i.e. $(" " if condition else None) --> $((" " if condition else None)) -- somewhat awkward but only done in ~2-3 infrequently used files, so seems to be worth it
• There are also two very interesting edge-cases that I left in the EXCLUDE_FILES for now -- /books/edit.html and /history/sources.html -- as both have HTML elements concatenated with +:
  

  openlibrary/openlibrary/templates/books/edit.html

  Line 30 in 3a59fc5

  note = '<p class="alert thanks">' + _("We already know about <b>%(work_title)s</b>, but not the <b>%(year)s %(publisher)s edition</b>. Thanks! Do you know any more about this edition?", work_title=work.title, year=edition.publish_date, publisher="; ".join(edition.publishers)) + "</p>"

  
  I'm not sure the best course of action for handling this would be (i.e. reverting to normal HTML, leaving them in the EXCLUDE_LIST, using the new escape syntax, something else, etc.)

Testing

1. pre-commit run --all-files with the current much shorter exclude list
2. You should see several warnings, all intentional to indicate purposefully escaped text, but the test should pass

Screenshot

Stakeholders

@cdrini @pidgezero-one

[pidgezero-one]

That's an interesting edge case - just my two cents, if concatenating HTML strings in that manner is infrequent/something we tend to avoid, I think it would be fine to just flag it with $# detect-missing-i18n-skip-line!

[rebecca-shoptaw]

That's an interesting edge case - just my two cents, if concatenating HTML strings in that manner is infrequent/something we tend to avoid, I think it would be fine to just flag it with $# detect-missing-i18n-skip-line!

@pidgezero-one I figure this will probably be the best solution, just a bit awkward for the files themselves as they each have many instances of the concatenated HTML, which will require many flags. But we'll see!

[pidgezero-one]

@rebecca-shoptaw Ahh, I see... It might be a little tricky, but I wonder if it's worth it to add some logic that detects if it's part of a $code block?

I guess HTML within $code blocks might still need i18n, so maybe even just ignoring tags that are followed by ' + _ - but I wonder if this isn't universal enough!

[mekarpeles]

We'll want to confirm that, in cases where we have dictionaries like this, that the values aren't being used as keys elsewhere in the code -- otherwise, once we i18n these keys, we could unintentionally break things. I don't think that's what's happening here, just sharing a possible use case we may hit given this PR touches some 100 files

[rebecca-shoptaw]

Roger that! I'll investigate -- and I believe this is the only dict affected by the PR (the script doesn't actually detect them, I just ran across it).

[rebecca-shoptaw]

Per @cdrini's advice, closing this to split it up into easier-to-review sub-PRs, which I will list and categorize in the original issue. 🙂