Wikisource Trusted Book Provider
- Titles in catalogue: ~500,000 in English alone!
- No real presence in OL yet; 60 works have wikisource IDs
- [x] Create identifier field for Wikisource
- Big Question: wikisource IDs are language specific (eg note the
enin https://en.wikisource.org/wiki/Goody_Two-Shoes_(1888) ) :grimacing: how should we deal with that? - Big Answer: Apparently URLs like this work! https://wikisource.org/wiki/en:The_Annotated_Strange_Case_of_Dr_Jekyll_and_Mr_Hyde
- Big Question: wikisource IDs are language specific (eg note the
- [ ] Display read button when provider ID present
- [ ] Custom download options
- Create separate issue for:
- Import missing records
- Set up automatic import pipeline
Proposal & Related Files
Follow the template in https://github.com/internetarchive/openlibrary/pull/8682/files ; except for wikisource!
FWIW, any item in Wikisource (and any other Wikimedia projects) willshould have a Wikidata item, so it might be enough to add the WD identifier and then get additional information (such as Wikimedia links) from that?
I don’t know enough about your code and data structures to know if that’s feasible for OL, but that’s generally how MusicBrainz and BookBrainz deals with e.g. getting Wikipedia excerpts.
Edit: It seems like not all do, though I believe it is the goal for all of them to have WD item… but I’m not familiar enough with Wikisource specifically to know if there’s maybe an exception there. So maybe either disregard this entire comment or consider allowing for Wikisource identifiers for cases where there’s no WD item. (That’s how we did in MB when WD was new and not all Wikipedia articles had WD items, but we knew that WD items for all WP articles was the goal.)
Some of the editors in Ukrainian Wikisource showed interest to this project. Specifically, to the potential traffic from this. I am slowly adding IDs to the Ukrainian editions. Hope this will soon be implemented
FWIW, any item in Wikisource (and any other Wikimedia projects) ~will~should have a Wikidata item, so it might be enough to add the WD identifier and then get additional information (such as Wikimedia links) from that?
I don’t know enough about your code and data structures to know if that’s feasible for OL, but that’s generally how MusicBrainz and BookBrainz deals with e.g. getting Wikipedia excerpts.
Edit: It seems like not all do, though I believe it is the goal for all of them to have WD item… but I’m not familiar enough with Wikisource specifically to know if there’s maybe an exception there. So maybe either disregard this entire comment or consider allowing for Wikisource identifiers for cases where there’s no WD item. (That’s how we did in MB when WD was new and not all Wikipedia articles had WD items, but we knew that WD items for all WP articles was the goal.)
@Freso Wikidata items are not very reliable for wikilinks. There is often conflation for Wikisource pages: one language page can point to a version page, which lists all the available editions of a certain work, while the other might be pointing to a specific edition.
I'll take this one on after our discussion earlier, @cdrini ! Thank you!
Awesome thank you @pidgezero-one !