Commit Graph

27847 Commits

Author SHA1 Message Date
Elian Doran
f441a145b5 fix(server): prod not starting due to bundling issues 2026-04-02 22:42:53 +03:00
Elian Doran
7189764916 chore(ocr): support overriding cache dir 2026-04-02 22:00:37 +03:00
Elian Doran
70bc707e3a chore(ocr): address requested changes 2026-04-02 21:58:54 +03:00
Elian Doran
90215bde8b chore(ocr): remove unnecessary index 2026-04-02 21:55:07 +03:00
Elian Doran
2b3ae5285b test(server): update integration DB to latest migration 2026-04-02 21:49:19 +03:00
Elian Doran
9b6d0db5b6 test(server): fix outdated tests in search result 2026-04-02 21:48:42 +03:00
Elian Doran
723da88ff8 chore(ocr): disable auto-processing by default 2026-04-02 21:46:05 +03:00
Elian Doran
5bcf2f4356 chore(deps): remove deprecated types for tesseract 2026-04-02 21:34:32 +03:00
Elian Doran
42680574c1 chore(deps): update pnpm lock 2026-04-02 21:34:03 +03:00
Elian Doran
82e723c915 test(ocr): fix broken tests 2026-04-02 21:27:46 +03:00
Elian Doran
32f95efa54 fix(ocr): image OCR in search results not shown 2026-04-02 21:14:56 +03:00
Elian Doran
3da416908d feat(ocr): display content snippet in quick search 2026-04-02 21:04:18 +03:00
Elian Doran
d79d2e9ad2 fix(ocr): too many blob queries in search 2026-04-02 20:58:11 +03:00
Elian Doran
30ba36894d chore(ocr): optimize search algorithm
OCRContentExpression now takes all tokens as an array (like NoteContentFulltextExp), iterates over the input note set from becca, and checks text representations in-memory — zero SQL queries.
parse.ts creates a single OCRContentExpression(tokens) instead of N separate instances.
The LIMIT 50 and the N+1 blob→note/attachment queries are gone entirely.
2026-04-02 20:54:22 +03:00
Elian Doran
b747402352 chore(ocr): get rid of costly ranking for OCR 2026-04-02 20:48:41 +03:00
Elian Doran
0398a9bda3 refactor(ocr): potential race condition with image upload 2026-04-02 20:40:17 +03:00
Elian Doran
72dff88384 refactor(ocr): get rid of unused routes and services 2026-04-02 20:34:37 +03:00
Elian Doran
0314a9755f refactor(ocr): minor changes 2026-04-02 20:32:58 +03:00
Elian Doran
bc967b15b2 chore(server): fix accidental changes 2026-04-02 20:28:17 +03:00
Elian Doran
8ac686a19f fix(ocr): TIFF overlapping with image processor 2026-04-02 20:26:31 +03:00
Elian Doran
aafecaa3a4 refactor(ocr): get rid of fake metadata 2026-04-02 20:24:31 +03:00
Elian Doran
bb23b08b15 refactor(ocr): get rid of unused clean up 2026-04-02 20:23:03 +03:00
Elian Doran
476396da53 refactor(ocr): deduplicate batch processing 2026-04-02 20:19:32 +03:00
Elian Doran
5112971848 refactor(ocr): reduce duplication 2026-04-02 20:17:24 +03:00
Elian Doran
2d852c38ec feat(ocr): automatic processing of attachments 2026-04-02 20:00:55 +03:00
Elian Doran
f163cacddc feat(ocr): integrate viewing attachment OCR 2026-04-02 19:51:11 +03:00
Elian Doran
6ecb1cb2b0 feat(settings): cross-reference OCR and language & region settings 2026-04-02 17:09:27 +03:00
Elian Doran
24fefe0711 refactor(ocr): remove unnecessary methods 2026-04-02 13:17:38 +03:00
Elian Doran
e5eba69d0d fix(ocr): cannot handle image/tiff 2026-04-02 12:51:58 +03:00
Elian Doran
bdd2b7e317 fix(ocr): properly handle office MIME types 2026-04-02 12:41:45 +03:00
Elian Doran
ad29375975 chore(ocr): remove unimplemented logic 2026-04-02 12:36:10 +03:00
Elian Doran
cf73a4ef43 feat(llm): integrate with OCR 2026-04-02 12:16:17 +03:00
Elian Doran
60a2621928 chore(ocr): remove last extraction date
Wasn't useful because blobs are hash-identified
2026-04-02 12:04:27 +03:00
Elian Doran
b4e5d9dbc2 feat(ocr): not well integrate with sync 2026-04-02 11:43:19 +03:00
Elian Doran
650b700415 feat(options/media): use a slider for JPEG quality 2026-04-02 11:17:54 +03:00
Elian Doran
6f2296eb05 feat(ocr): use a slider for confidence 2026-04-02 11:09:36 +03:00
Elian Doran
722efd74c2 fix(ocr): default confidence level is too low 2026-04-02 11:06:58 +03:00
Elian Doran
5dc9b6defe chore(ocr): deduplicate & fix percentage for confidence in log 2026-04-02 11:04:26 +03:00
Elian Doran
605fbaaa4a fix(ocr): automatic OCR not respecting language 2026-04-02 11:01:20 +03:00
Elian Doran
23b46865c5 refactor(ocr): simplify initialization of image processor 2026-04-02 10:59:58 +03:00
Elian Doran
ac310eaaf5 feat(ocr): handle cache dir properly 2026-04-02 10:54:15 +03:00
Elian Doran
010f59df8a chore(ocr): make OCR text representation selectable 2026-04-02 10:25:41 +03:00
Elian Doran
44a5dccd61 chore(ocr): remove master switch 2026-04-02 10:22:34 +03:00
Elian Doran
acbbf021a1 refactor(ocr): remove unnecessary translations 2026-04-02 10:13:03 +03:00
Elian Doran
731fece258 feat(ocr): reintroduce batch processing 2026-04-02 10:08:24 +03:00
Elian Doran
8d255d1b89 feat(ocr): make "process OCR" always reprocess 2026-04-02 10:02:06 +03:00
Elian Doran
64318c92e7 fix(ocr): route default interfering with content language 2026-04-02 10:00:12 +03:00
Elian Doran
49fc7e48d4 feat(ocr): integrate with content language 2026-04-02 09:52:28 +03:00
Elian Doran
ec9fa0baee chore(options): rename options to match media scope 2026-04-01 22:42:17 +03:00
Elian Doran
ba91d91fd1 chore(options): start adding options for OCR 2026-04-01 22:37:32 +03:00