OPDS enchancement

[IvanGavrilenko]

What would you like to improve?

All these enchancements are only for OPDS engine, not for ordinary web view

1. Translate OPDS folders to the web-interface language
2. Allow an number of maximum amount on Author or Series page (or at least allow such feature - better on server level, not on opds user level)
3. Show and sort Authors in the "last-name first-name" order (or at least allow such feature - better on server level, not on opds user level)
4. Enrich Author page with books-by-series (then leading to the series name folder structure), books-without-series, alphabeticaly-sorted-books and by-date-added books
5. Enrich book info page - add hyperlinks leading to the all-authors-books and all-series opds books list

What's the current friction?

1. All folders are only in English
2. There is one list containing all authors (currently I have list of 200+ authors)
3. Currently authors are shown as first-name last-name sorting (i.e. Lewis Carroll)
4. Currently as you enter the author's page you see all books ordered by the OPDS principle
5. The book info page does not have these links

What would make this better?

1. Just a cosmetics update
2. If you have more than 20 authors it's not convenient to scroll down looking for a specific author or Series name but if authors are splitter to S -> Sh -> Shakespeare (and the page with Shakespeare has not more than 30 or other specified number authors) the time would be taken much less
3. The first-name last-name sorting is normal for web UI as I can quickly type the needed author to the search line but for e-ink reader it is not so convenient - more surely I would remember the authors last-name than the first-name and searching is more long due to the e-ink response time, so looking for the "Carroll, Lewis" instead of "Lewis Carroll" can be time saving
4. When an author has 3 series and 5 stand alone books and I remember just the first book of the series I can not download the entire series quickly - I have to open the remembered book, look for the series name, go to series page, look for the series and then download the entire series. While the following structure would allow to find all books much quicker:

• Author

  -by series

  -series1
  
      -book1
      -book2
  
  -series2
  
      -book3
      -book4
  

  -not in series

  -book5
  -book6
  

  -all books by name

  -book1
  -book2
  -book3
  -book4
  -book5
  -book6
  

  -all books by date added

  -book1
  -book5
  -book3
  -book6
  -book4
  -book2
  

5. Currently if I open the book I can not find all books of the author or all books in the series quickly (look to the p.4 explanation). And if series contain books of several authors (i.e. https://en.wikipedia.org/wiki/The_Three_Investigators) it's hard to download all books

Anything else? (Optional)

2. It seems to me that Calibre has the similar feature

The 2-5 enchancement examples might be found in https://flibusta.is/opds (most famous fb2 library in Russian language)

Before Submitting

• I've searched existing issues and confirmed this enhancement hasn't been requested yet

[eLindros]

I'd like also to suggest that one can have different OPDS-Feeds for each library. Currently I have a library for ebooks and one for comics. I use different apps to consume the books in those libraries (Readest and Panels on iOS). So it's kinda annoying to scroll through every comic series when looking for a book an vice versa. The api could be extended to /api/v1/opds/{library} for that.

With that I coud stop using Komga just for comics and move everything into Grimmory.

[IvanGavrilenko]

It seems to me that current realization of OPDS has Library folders, but of cource if you have more than 20 books in each library - it's not convenient to look for a book

[eLindros]

Yes, there are library folders. But my comic selection is large and organized by series. That's why I thought to just separate the whole opds feeds.

[eLindros]

Nevermind,I switched to ComicOPDS which has all what I needed. I keep my comics separate from my ebooks for now.

[SergeiAK]

Big +1 on all of these — I'd been meaning to ask for several of them myself. A few notes from digging through the OPDS source, in case it helps gauge effort:

#1 (translate OPDS folder names) — agree this should be easy and it bugs me too. The folder titles ("Authors", "Series", "Recent", …) appear to be hardcoded English strings in the feed-generation code; moving them to the same i18n/message-bundle mechanism the web UI already uses would cover it. One small wrinkle: an OPDS reader doesn't necessarily report its language the way the browser does, so the feed language is probably best tied to a setting (server-level or per OPDS user) rather than guessed from the client.

#3 ("Last name, First name" author sort) — strong +1, this is a big one for non-English libraries. Mine is mostly Russian, and on an e-ink reader you remember an author by surname and search is slow, so browsing/searching "Толстой, Лев" instead of "Лев Толстой" genuinely saves time. Calibre handles exactly this with a separate author-sort field ("Last, First") distinct from the display name, which might be a useful model. Agree with you that for #2/#3 a server-level setting feels right; for reference, a per-OPDS-user flag could also ride on the existing OpdsUserV2 + PATCH /api/v2/opds-users/{id} with no new endpoints, so either convention works — just worth picking one for all these toggles.

#4 (group an author's books by series) — this is the one I care about most, and it looks cheap on top of what's already there:

• OpdsFeedService.generateAuthorsNavigation(...) already builds a navigation feed, but each author entry links straight to the flat acquisition feed /api/v1/opds/catalog?author=<name>. The author→series view is one more navigation level at that point.
• The data is already on hand: generateCatalogFeed already fetches an author's books via opdsBookService.getBooksByAuthorName(...), and every book entry already emits seriesName / seriesNumber (the belongs-to-collection + group-position metadata). So "by series" is just grouping the per-author result by seriesName in memory — no new queries or schema.
• The navigation-feed pattern already exists twice (generateAuthorsNavigation / generateSeriesNavigation), so the new feed can reuse it plus the existing escapeXml / URLEncoder helpers. The "all books by name / by date added" sub-lists map onto the existing applySortOrder — just different links from the same author node.

Two decisions worth making up front for #4:

1. Series sort order. The existing /catalog?series=<name> feed runs through applySortOrder (default RECENT), so it is not ordered by series index. For correct reading order the series-scoped list needs explicit sorting by seriesNumber (1, 1.5, 2, …).
2. Series shared across authors (your Three Investigators example in chore(deps): bump the gradle-dependencies group in /booklore-api with 5 updates #5). generateCatalogFeed currently treats author and series as mutually exclusive (if author … else if series …), and there's no combined query — only getBooksByAuthorName / getBooksBySeriesName. Scoping a series to the current author would need a small composite filter (?author=X&series=Y + one new query method).

Happy to write #4 up as a focused enhancement issue with the proposed feed structure if a maintainer wants to split it out.