v13.2.0

New

• enh: Improve self-documentation in the admin settings
• perf(dav): Load detections in bulk
• perf: Only run userMountCache->clear() per batch

Fixed

• fix(RemoveDuplicateFaceDetections): Prioritize clustered faces + work in batches
• fix: Prevent duplicates and clean up remaining near duplicates
• fix(FsMove): Remove unused columns
• fix(insertMove): Use a transaction
• fix(onAccessUpdate): Correct user access for shared subdirectories
• fix(FsActionService): Don't lose subdirectory files in onMove
• fix(FileListener): Queue move actions correctly and fix FsActionMapper
• fix(dav): Check ownership of face detections
• fix(dav): Only return the current user's face detections
• fix(ClusterFacesJob): Add a global limit on clustering runs
• fix(ClusterFacesJob): Calculate clustering memory correctly and use a lock per user
• fix(FsActionMapper): Make sure to use DB index
• fix: Remove google-img-scrap
• fix: log model download URL and destination (thanks to Ravi Shankar)
• chore(deps): Update dependencies
• fix(l10n): Update translations from Transifex

v12.0.3

Fixed

• fix(FsMove): Remove unused columns
• fix(insertMove): Use a transaction
• perf: Only run userMountCache->clear() per batch
• fix(onAccessUpdate): Correct user access for shared subdirectories
• fix(FsActionService): Don't lose subdirectory files in onMove
• fix(FileListener): Queue move actions correctly and fix FsActionMapper
• fix(dav): Only return the current user's face detections
• fix(dav): Check ownership of face detections
• fix(FsActionMapper): Make sure to use DB index
• fix: Remove google-img-scrap

v13.1.0

New

• feat(deps): Add Nextcloud 36 support

Fixed

• fix(l10n): Update translations from Transifex
• fix(nodejs): Upgrade shipped nodejs to v22
• fix(ClusterFacesJob): Adjust clustering badge size to memory limit automatically
• fix(FsActionService): Schedule ClusterFacesJob with userId casted to strings
• fix: Fix admin page

v12.0.2

Fixed

• fix: Upgrade node version

v12.0.1

Fixed

• fix(ClusterFacesJob): Adjust clustering badge size to memory limit automatically
• fix(FsActionService): Schedule ClusterFacesJob with userId casted to strings
• fix: Fix admin page

v13.0.0

Breaking changes

• Drop support for nc < 35

New

• feat: Add recognize_backend exApp integration: Payloads can now be processed on a separate machine via AppAPI
• feat(deps): Add Nextcloud 35 support

Fixed

• fix(QueueMapper): Do not mutate queueItem
• fix: inject HTTP proxy env vars into node subprocess calls (thanks to @likehopper)
• chore: Upgrade to vue 3 and nextcloud-vue 9
• fix(l10n): Update translations from Transifex

v12.0.0

Breaking changes

• Drop support for Nextcloud 33

Fixed

• fix: Add migration changing non-lazy config entries to lazy
• fix(l10n): Update translations from Transifex
• perf: Port away from more getById
• fix: Migrate getById to getFirstNodeById
• fix: do not declare the id column type
• fix: Do not attempt to classify TIFF files

v11.0.1

Fixed

• fix(l10n): Update translations from Transifex
• refactor(Application): Use modern SabrePluginAddEvent
• fix(FileListener): Do not consider events in trashbin
• refactor: Process all fs events asynchronously
• fix(ffmpeg): Autodetect system installed ffmpeg and allow manually setting ffmpeg path
• fix(ClusteringFaceClassifier): Make sure to schedule clustering job
• fix: Improve error handling

v10.0.7

Fixed

• fix(ClusteringFaceClassifier): Make sure to schedule clustering job

v9.0.9

Fixed

• fix: Don't leak global functions in php scoper