A shared test_support module in db.rs provides the fixture: sqlite-vec
registered via auto-extension, an in-memory connection with foreign keys
on, and both migrations applied - no refactoring of production code
needed since every query function already takes &Connection.
db.rs coverage: folder idempotency, upsert_image update semantics
(favorite/rating preserved, AI tag state invalidated), the get_images
filter matrix with pagination and count_images agreement, tag
merge/rename/delete, user-tag precedence over AI tags in update_ai_tags,
album CRUD with FK cascade, the embedding job queue (backfill, retry,
consistency repair), tag search, and delete_folder cascades.
vector.rs coverage: pack/unpack round-trip, embedding upsert/delete with
dimension validation, and find_similar_image_ids ranking on both the
global KNN and folder-scoped brute-force paths.
The visual k-means cluster feature was confusingly named tag_cloud / tagCloud /
TagCloud across the whole stack, while the actual tag list is explore_tags — the
two were trivially easy to mix up (and did cause confusion). Rename the cluster
side to visual_cluster / visualCluster / VisualCluster everywhere: command
get_tag_cloud -> get_visual_clusters (+ lib.rs registration and the invoke
string), VisualClusterEntry, the store fields/actions/tokens, and the mock
backend. Old names are retired rather than reused, so any missed reference fails
loudly instead of silently resolving to the wrong concept.
The tags side keeps its accurate explore_tags naming, and the user-facing
"Tag Cloud" UI label is unchanged.
Also rename the SQLite tag_cloud_cache table -> visual_cluster_cache (the old
table is dropped during schema setup — it is a disposable cache already
invalidated by the clustering version bump) and the TagCloud.tsx component file
-> ExploreView.tsx, since it is the Explore container hosting both the cluster
and tag views.
The get_tag_cloud cache key was built by loading and hashing every embedding
blob for the scope *before* checking the cache, so even a cache hit re-read
hundreds of MB on large libraries and stalled Explore for several seconds.
Validate the cache from a lightweight image-ID-set signature plus the embedding
revision instead, so a hit never loads embeddings. The ID-set hash keeps the key
membership-sensitive (add/remove/move between folders) and the revision covers
an image being re-embedded in place. Cache write failures are now logged rather
than silently ignored.
On the frontend, switching folders (or re-entering Explore) no longer leaves the
previous folder's clusters/tags on screen with no loading indicator:
loadTagCloud/loadExploreTags clear stale entries on a real folder switch. The
displayed folder is tracked separately (exploreTagsShownFolderId) from the
cache-dirty marker so a same-folder invalidation (tag edits, new AI tags) does
not masquerade as a switch and wipe the visible list mid-refresh.
Similar search scoping:
- Add current_album as a similar-scope option and remember the source album when similar or region searches are launched from an album.
- Route gallery and lightbox similar actions through scope-aware store helpers so Album/Folder/All choices are applied consistently.
- Keep pagination and scope toggles working for both whole-image and region-search result sets.
Backend filtering:
- Extend find_similar_images and find_similar_by_region params with album_id, giving album scope precedence over folder scope.
- Add album_membership filtering for HNSW whole-image search and brute-force crop embedding search.
Drops and recreates the sqlite-vec tables at the current model dimension, then
re-queues every image for embedding. Fixes "dimension mismatch" search errors
that occur when the vector table was built for a different model/dimension (e.g.
after experimenting with 768-dim models against this 512-dim build).
The rebuild runs under the embedding worker's DB write lock and resets the job
queue inside a single transaction, so it can't interleave with an in-flight
embedding batch (per code review).
Auto-fix pass (uninlined format args, needless borrows, to_vec,
size_of_val, map_err->inspect_err) plus hand fixes: fully annotated the
sqlite-vec transmute (c_char for cross-platform correctness), type alias
for the duplicate-scan tuple, slice signatures over &mut Vec/&PathBuf,
and #[allow(dead_code)] documenting the intentionally dormant caption
worker. The CI clippy step now fails on any new warning.
Adds a full discovery and organisation layer to the gallery.
## AI Tagger
- WD tagger (ONNX via ort) with DirectML/CPU acceleration
- Per-image and per-folder tag queuing; configurable batch size and
confidence threshold; model downloaded on first use via ureq/zip
- Tags stored with source ('ai' | 'user'); user tags are never
overwritten by AI; AI tags protected from accidental promotion
- Tagger pause/resume per folder; in-flight batches discarded cleanly
on pause or cancellation without leaving jobs stuck in processing
## Semantic & Tag Search
- CLIP text-query embedding via candle (HuggingFace hub model)
- Progressive candidate doubling with filter post-processing so
folder/rating/media-kind filters do not silently under-return results
- Tag search with DB-backed pagination (offset + COUNT total)
- Filename search unchanged; prefix syntax: s:, t:, f: in search bar
## Similarity Search
- HNSW in-memory index (hnsw_rs) with monotonic embedding_revision
counter for cache invalidation; build_index retries if a concurrent
write advances the revision during construction
- Folder-scoped similar search uses brute-force cosine scan (no k limit)
- Region-based similarity: crop an arbitrary rectangle in the lightbox,
embed it with CLIP, find the nearest images globally or within folder
- Pagination for both similar and region results; scope toggle
(all media / current folder) re-runs the query without reopening image
## Duplicate Finder
- Three-phase detection: stat (size) → sample hash (4×16 KB windows)
→ full-file hash (xxh3, large files only) to eliminate false positives
- Filesystem-first deletion: only removes DB rows for files successfully
deleted from disk; failed files remain visible for retry
- Persisted scan cache (SQLite) with invalidation on reindex and deletion;
both global and per-folder scopes invalidated after any deletion
## Tag Cloud & Explore
- Visual cluster explore: k-means over CLIP embeddings, configurable k,
representative thumbnail per cluster; cache keyed on xxh3 of embedding set
- Tag explore: ranked tag list with image counts and representative
thumbnail per tag; invalidated when AI tagging completes or folder removed
- Tag autocomplete for search bar
## Folder Management
- Inline rename (display name only, not OS rename)
- Relocate: file-picker to update folder path; all image paths rewritten
using SUBSTR prefix replacement to avoid corrupting paths that contain
the folder name as a substring
- Missing-folder recovery banner: Locate or Remove, never silent deletion
- Right-click context menu on sidebar items: Reindex, Rename, Locate,
Remove; hover buttons preserved alongside context menu
## Background Tasks
- Per-folder progress panel with embedding, tagging, caption, and
thumbnail stage breakdown
- Retry button per task for failed embeddings and tagging jobs
- Completed-task notifications (system tray via tauri-plugin-notification)
- Scan errors shown inline; WalkDir permission errors protect existing
records from deletion rather than cascading data loss
## Backend correctness
- sqlite-vec DML performed outside transactions throughout (virtual-table
limitation); embedding revision incremented on delete as well as insert
- Tagging job discard checks status = 'processing' so paused jobs reset
to 'pending' are also skipped, not just cancelled ones
- Stale async responses in Lightbox guarded with cancellation flags and
currentImageIdRef for both tag-load and tag-add callbacks
- Duplicate cache cleared on reindex; folder-removal clears vector rows
outside transaction before deleting the folder row
- Fix root cause: embedding_source_path() returns Result<PathBuf>, returning
Err for videos without a thumbnail instead of silently falling back to the
raw .mp4 path that CLIP cannot decode
- indexer: split embedding batch into pre-failed (no source) and embeddable
jobs; pre-failed are marked immediately without hitting the CLIP model
- db: retry_failed_embedding_jobs skips videos still without a thumbnail so
they no longer re-fail immediately on retry
- Add get_failed_embedding_images command listing failed files + error per folder
- Gallery: amber warning badge on tiles with embedding_status = 'failed'
- BackgroundTasks: fetch and show failed filenames/errors in expanded panel
- Toolbar: conditional 'Failed Embeddings' amber filter pill shown when any
folder has embedding_failed > 0; filters at DB level via new
embedding_failed_only param on get_images / count_images
- TagCloud: replaced vocabulary/dictionary label system with representative
image thumbnails per cluster; results cached in SQLite by image-id hash
Introduces an Explore view with a tag cloud that clusters image embeddings
using cosine k-means and labels clusters via vocabulary-nearest-neighbour
CLIP matching. Vocabulary embeddings are disk-cached (FNV hash-keyed) to
avoid redundant inference. Enables CUDA for candle dependencies and adds a
build.rs check that surfaces a clear error when the toolkit is missing.
- add semantic text search with toolbar mode switching and sqlite-vec query support
- improve embedding progress visibility, failure recovery, and similar-image affordances
- add search clearing and keyboard controls for filename vs semantic search modes
- refine background task interactions and gallery/lightbox embedding states
Refs: #4
- add Candle + HF Hub CLIP image embedding pipeline with background embedding worker
- write image embeddings into sqlite-vec and expose similar-image lookup through a new backend command
- surface embedding progress and recovery in the UI, including retries for failed embeddings
- improve gallery/lightbox embedding UX and make similar-image actions directly accessible
Refs: #3, #4
Sets up Tauri v2 + React frontend with SQLite metadata store,
sqlite-vec for CLIP embedding infrastructure, and r2d2 connection
pool replacing the single Arc<Mutex<Connection>>. Indexer now uses
rayon for parallel file metadata collection.
LyAhn