* feat: add OCR functionality and related configurations
* chore: update labeler configuration for machine learning files
* feat(i18n): enhance OCR model descriptions and add orientation classification and unwarping features
* chore: update Dockerfile to include ccache for improved build performance
* feat(ocr): enhance OCR model configuration with orientation classification and unwarping options, update PaddleOCR integration, and improve response structure
* refactor(ocr): remove OCR_CLEANUP job from enum and type definitions
* refactor(ocr): remove obsolete OCR entity and migration files, and update asset job status and schema to accommodate new OCR table structure
* refactor(ocr): update OCR schema and response structure to use individual coordinates instead of bounding box, and adjust related service and repository files
* feat: enhance OCR configuration and functionality
- Updated OCR settings to include minimum detection box score, minimum detection score, and minimum recognition score.
- Refactored PaddleOCRecognizer to utilize new scoring parameters.
- Introduced new database tables for asset OCR data and search functionality.
- Modified related services and repositories to support the new OCR features.
- Updated translations for improved clarity in settings UI.
* sql changes
* use rapidocr
* change dto
* update web
* update lock
* update api
* store positions as normalized floats
* match column order in db
* update admin ui settings descriptions
fix max resolution key
set min threshold to 0.1
fix bind
* apply config correctly, adjust defaults
* unnecessary model type
* unnecessary sources
* fix(ocr): switch RapidOCR lang type from LangDet to LangRec
* fix(ocr): expose lang_type (LangRec.CH) and font_path on OcrOptions for RapidOCR
* fix(ocr): make OCR text search case- and accent-insensitive using ILIKE + unaccent
* fix(ocr): add OCR search fields
* fix: Add OCR database migration and update ML prediction logic.
* trigrams are already case insensitive
* add tests
* format
* update migrations
* wrong uuid function
* linting
* maybe fix medium tests
* formatting
* fix weblate check
* openapi
* sql
* minor fixes
* maybe fix medium tests part 2
* passing medium tests
* format web
* readd sql
* format dart
* disabled in e2e
* chore: translation ordering
---------
Co-authored-by: mertalev <101130780+mertalev@users.noreply.github.com>
Co-authored-by: Alex Tran <alex.tran1502@gmail.com>
* Enable filteing by example
* Drop `@GenerateSql` for `getEmbedding`?
* Improve error message
* PR Feedback
* Sort en.json
* Add SQL
* Fix lint
* Drop test that is no longer valid
* Fix i18n file sorting
* Fix TS error
* Add a `requireAccess` before pulling the embedding
* Fix decorators
* Run `make open-api`
---------
Co-authored-by: Alex <alex.tran1502@gmail.com>
The height of the search results element was unrestricted, which meant that the
asset visibility calculations were completely incorrect. The consequence of
this is that assets which should not have been visible, were. In practical
terms, all assets below the viewport were rendered when they shouldn't have
been which is terrible for performance. Limiting the height of the viewport
fixes that calculation and assets are correctly hidden.
The consequence of limiting the height of the viewport is that the intersector
then incorrectly thought the scroll position was always at the end. This has
been fixed by calculating the position of sliding window against the calculated
asset layout container height.
Co-authored-by: Alex <alex.tran1502@gmail.com>
The contents of search results are slightly offset by the search bar, search
terms and spacing (margins/padding), and needs to be factored in when
calculating whether an asset is visible or not. The offset was 0, which
meant that assets were removed from view too early.
The search results page can become unstable with large amounts of assets, and
has therefore been limited to displaying just 5000 assets. This limit is
arbitrary and may be too restrictive.
