Show editions in reading log

[benbdeitch]

Closes #2733

This PR alters how the 'already-read' and 'currently-read' bookshelves retrieve and display items. Rather than simply fetching the work data, they now retrieve the edition-specific data that is already stored in their reading log.

Currently, the 'want-to-read' bookshelf is untouched, as users might want to know when any edition becomes available to borrow.

Technical

The get_users_logged_books function in openlibrary/core/Bookshelves was edited to provide this functionality. Based on the bookshelf_id received, it switches on editions variable, that is used to later define behavior.

Both the get_sorted_reading_log_books and get_filtered_reading_log_books were altered to switch between fetching work data, or fetching work and edition data, based on the 'editions' variable. As well, a new function, link_editions_to_works was created to process the retrieved results, and properly tie them together, so that they are correctly interpreted by the SearchResultsWork macro.

Testing

To verify, simply add a few books to each shelf of your reading log. Preferably, editions with titles that differ from the work's should be chosen, or a brief edit in macro/SearchResultsWork can be used to display the doc_type variable. On each shelf, one should both check the shelf as a whole (which uses the 'get_sorted_reading_log_books' function, and attempt to search for an individual work (using the get_filtered_reading_log_books function).

Screenshot

(As noted, the 'solr_edition' line is not a part of the pull-request; merely an easy way to display the doc_type variable for testing.)

Stakeholders

@cdrini

[Watt3r]

I bump into this problem often and this would totally fix it for me. Thanks for implementing this @benbdeitch!

[RayBB]

@benbdeitch how is it going with this? Do you need any help?

[cdrini]

This is looking great! A few code cleanups, and one potential error! Will take a look at the last method afterwards.

Testing:

This page seems to error: https://testing.openlibrary.org/people/bigdbag/books/already-read?page=10&debug=true

✅ sort | edition_id=null | https://testing.openlibrary.org/people/pyat/books/already-read
✅ filter | edition_id=null | https://testing.openlibrary.org/people/pyat/books/already-read?q=playground
✅ sort | large reading log | https://testing.openlibrary.org/people/bigdbag/books/already-read
✅ filter | large reading log | https://testing.openlibrary.org/people/bigdbag/books/already-read?q=harry+potter

[cdrini]

Here's the finished review!

[cdrini]

Performance-wise, so far this is looking good!

Will testing the filtered case once the bug is fixed causing it to return 0 results!

[cdrini]

Ok the bug was with ol-solr1 , so not related to this code! Performance is looking good as well:

I'll do a few more tests later today, and then I think this is good to merge!!

[cdrini]

Ok lgtm! I tested redirects; there is one edge case when you add two works and then the two works are merged... but I think it's very edge case and I'm not too sure how we should handle it. It does work correctly when you add a work/edition and then that work gets merged into another. Also works if you add just a work, and it gets merged into another -- displays the empty tile it currently displays.