Removes the OpenCLIP-on-ONNX classifier and everything that fed or
consumed it:
- backend: app/services/vision/, app/tasks/vision.py,
app/services/feature_flags.py, app/routers/features.py — all
deleted; admin AI/feature-flag endpoints and the worker-vision
bootstrap call gone. Photo.needs_review and its index dropped.
- frontend: AI Settings tab, useFeaturesQuery hook, FeatureFlag
types, "Needs Review" sidebar entry + filter, needs_review filter
URL param all gone.
- infra: worker-vision compose service + models_data volume deleted;
worker-light command no longer runs bootstrap_models; the db
image switches from pgvector/pgvector:pg16 to postgres:16; backend
Dockerfile drops the dedicated torch RUN layer; requirements.txt
drops torch/torchvision/open-clip-torch/onnxruntime.
Alembic 0019_drop_ai_remnants:
- drops photos.needs_review + ix_photos_needs_review
- DROP EXTENSION IF EXISTS vector (must run before the image swap;
the new postgres:16 doesn't ship pgvector)
New scripts/full_refresh.py: one-shot DB ↔ filesystem reconciliation.
Runs cleanup_data_integrity, scans every active SourceRoot inline
(no celery dependency so the worker can be stopped), hard-prunes
photo + folder rows for files that are gone, removes orphan
/data/thumbs/{user}/{photo}/ directories. New helper
prune_orphan_thumbnails in cleanup.py.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The /playback transcode cache lives at /data/video-cache. That
directory was created in container-local storage (the mkdir at
services/video.py import time), not on a shared volume — so the
worker pretranscode populated its own ephemeral copy while the
backend served from a different empty one, and the cache evaporated
on every container restart.
Add a video_cache_data named volume mounted on backend, worker-light,
and worker-vision so the pretranscoded files actually reach the
serving path and survive deploys.
The workers couldn't decrypt users.nextcloud_app_password_enc because
SECRET_KEY wasn't in their env. _credentials_for() then raised
NextcloudCredentialsMissing and our code swallowed it as "no NC
auth → fall back to local path."
Surfaced on the Phase 4 deploy when /data/thumbs/.../medium.webp
was purged and the vision worker had no disk fallback left. NC
preview fetch then returned None, the classifier got no image,
and the photo failed to classify.
Also masked Phase 3 silently — extract_metadata in worker-light
was falling back to ExifTool every time instead of hitting Memories
(which would have been fine because ExifTool produces the same
fields, but slower and unnecessary). With SECRET_KEY available, the
Memories primary path actually fires.
End-to-end webhook flow is proven on this NC instance (NodeCreated +
NodeWritten both fired and dispatched scan_folder on a PUT test), so
the watchfiles-based polling layer is no longer needed.
- scanner.start_initial_scan no longer queues watch_folders on boot.
- scan.watch_folders kept as a one-line no-op shim so any leftover
apply_async in flight from the previous deploy doesn't crash a
worker. Will be deleted entirely after the queue drains.
- celery.py reroutes watch_folders to the `default` queue (worker-light)
so the no-op shim actually completes — the `watcher` queue is dead.
- docker-compose drops the mulita-worker-watcher service. Its celery
--beat responsibility (firing discard_missing_photos_beat every 30
min) moves to worker-light's command.
Latency note: NC dispatches webhook events through its background-job
queue, currently run by cron */5. After this commit lands you'll want
to tighten cron to */1 so new uploads land in mule within ~60s instead
of up to 5 min.
The watchfiles-based watcher works but duplicates Nextcloud's own
notion of "this file changed." NC has a webhook_listeners app that
can POST file events to an external URL. This adds the mule side of
that handshake.
- POST /api/v1/internal/nc-webhook authenticates a Bearer token
(NEXTCLOUD_WEBHOOK_SECRET, hmac.compare_digest) and dispatches the
same scan_folder / handle_file_deletion machinery the watcher used.
- Handles NodeCreated, NodeWritten, NodeDeleted, NodeRenamed.
Renamed is mapped to delete-old + scan-new-parent. Maps NC's
/admin/files/... path to the bind-mounted /nextcloud-users/admin/files/...
- backend/scripts/register_nc_webhooks.py is the idempotent
registrar: lists existing webhooks, deletes any pointing at the
target URL, then POSTs four fresh ones via OCS.
- Sets the env passthrough on backend + all workers in compose so
the same secret is available wherever the registrar might run.
watch_folders stays in place for now — webhooks become primary, the
watcher is a belt-and-suspenders fallback. Drop the watcher in a
follow-up once webhooks are proven reliable on this NC instance.
Two related fixes for the Nextcloud library lifecycle.
1. DELETE /api/v1/nextcloud/source-roots/{id} now actually deletes
the SourceRoot, every Folder under it, and every Photo in those
folders (Nextcloud files untouched). Was a soft-deactivate
(is_active=false) that left the rows around forever, so re-adding
the same path resurrected ghosts and prune-missing reported zero.
Returns {deleted_photos, deleted_folders}; the Settings UI toasts
the count and invalidates photos/folders/stats so cached lists
don't show ghosts. photo_tags and heap_photos already cascade via
ON DELETE CASCADE; FolderShare uses a stringly-typed folder_id
with no FK so cleaned up explicitly.
2. The watcher (watch_folders task) was getting killed every five
minutes by the global task_soft_time_limit=300 in app/tasks/celery.py
despite passing soft_time_limit=None on the decorator (None falls
back to the worker default in this Celery version). Override with
soft_time_limit=0, time_limit=0 (= unlimited) so the watch loop
actually stays alive. The 'Soft time limit (300s) exceeded' /
'Worker exited prematurely' lines should stop in worker-watcher
logs.
3. Added discard_missing_photos() in services/cleanup.py — a soft
variant of prune_missing_photos that walks every present source
root, checks os.path.exists for each non-discarded Photo, and
flips is_discarded=true on the missing ones (UPDATE not DELETE).
Wired as discard_missing_photos_beat in tasks/scan.py and
scheduled every 30 min via celery beat. Beat runs in-process on
worker-watcher (--beat flag in compose) — there's only ever one
watcher and we don't need a separate container.
Hard delete remains manual via prune-missing for users who want to
review before committing. The beat catch-up only soft-discards (file
gone -> mule-image trash, restorable).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Production runs were on the dev --reload single-worker config. The frontend
fans out ~15 parallel API calls on first paint (folders/tree, tags, heaps,
sharing/*, stats, photos, worker-status, scan/status); they all serialized
on one event loop and felt slow. Switch to 2 workers without --reload for
real concurrency. --proxy-headers preserved client IPs through nginx.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Without this, the docker default resolver forwards the lookup to the
host gateway, which returns the public IONOS VPS IP. cloud.hubris is
not in the VPS traefik exposure list, so TLS handshakes during
WebDAV calls die with httpx.ConnectError: SSL UNEXPECTED_EOF.
extra_hosts pins it to caddy on 192.168.8.175, which holds the
cloud.hubris.network cert and proxies to the Nextcloud LXC. Applied
to every service for symmetry; only backend currently makes the
WebDAV calls.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Lets each mule-image user (matched via OIDC preferred_username,
overridable in Settings) browse their Nextcloud files/ tree from the
mule-image UI and register subfolders as per-user SourceRoots. Reads
stay direct on the bind-mounted /nextcloud-users path; mutations
(upload, delete, rename, move within NC) dispatch through Nextcloud
WebDAV so oc_filecache, trashbin, comments, and desktop-sync clients
stay coherent.
Backend:
- users.nextcloud_username + nextcloud_app_password_enc (Fernet at rest,
key derived from SECRET_KEY) — alembic 0016
- services/nextcloud_dav.py: minimal WebDAV client (PUT, MKCOL, DELETE,
MOVE) with HTTP Basic auth via the per-user app password
- routers/nextcloud.py: GET /browse, /whoami, GET/POST/DELETE
/source-roots (path-scoped to current_user.nextcloud_username with
realpath traversal guard)
- PATCH /api/v1/auth/me to update nextcloud_username and app password
- OIDC callback defaults nextcloud_username from preferred_username on
first login; backfill on existing users; never overwrites a manual
override
- routers/upload.py: stream upload to NamedTemporaryFile, then PUT to
WebDAV (with MKCOL chain) when destination is NC-rooted; existing
Photo row creation runs unchanged
- routers/discard.py empty-trash: WebDAV DELETE for NC files
- routers/photos.py rename + move: WebDAV MOVE for NC paths;
cross-system move/copy returns a clean error
- routers/folders.py rename + create + permanent-delete: dispatch via
WebDAV when targeting NC-rooted paths
Frontend:
- AuthUser carries nextcloud_username + has_nextcloud_app_password
- services/api.ts: nextcloud + account namespaces
- components/dialogs/NextcloudFolderPicker.tsx: lazy tree browser, name
+ submit -> POST /source-roots
- SettingsDialog: new "Nextcloud library" card with username override +
validate, app-password input, list/remove of NC libraries, and the
picker entry point
docker-compose.yml: NEXTCLOUD_USERS_HOST_PATH bind to /nextcloud-users
on backend + 3 workers; NEXTCLOUD_USERS_ROOT + NEXTCLOUD_BASE_URL env.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds OIDC_LINK_BY_USERNAME as a last-resort linking step after
(issuer, sub) and email both miss. Matches IdP preferred_username
against users.username.
Why: local accounts created before OIDC never collected an email
(no UI for it), so the email fallback cannot relink them. A new
SSO login therefore falls into JIT and creates username-1. On a
single-tenant homelab where the IdP owns the namespace, matching
by username is safe and makes first-time SSO transparent for
pre-existing users. Gated behind a flag so multi-tenant deployments
keep the stricter default.
Adds optional SSO via Authentik (or any OIDC provider) alongside the
existing password flow, and pulls profile images from the provider's
`picture` claim or Gravatar so the sharing UI stops looking anonymous.
Password login stays available as a recovery path; JIT provisioning and
admin-group mapping are env-configurable.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Addresses 16 robustness, transparency, and performance issues across
the Celery media processing pipeline:
Critical:
- Singleton DB engine in vision tasks (was leaking one per task call)
- acks_late + task_reject_on_worker_lost so crashed workers don't lose tasks
- Global soft/hard time limits (5/10 min) to prevent hung worker slots
- Thumbnail copy-before-resize (in-place mutation degraded larger sizes)
- backfill_vision now checks each task type independently (OCR, faces, etc.)
- Parameterized LIMIT in backfill_vision (was f-string SQL injection)
High:
- try/except + retry(max=3) on all vision inference tasks
- extract_metadata writes processing_error on exiftool failure
- PIL Image handles closed in _load_thumb/_load_original
- Scan progress Redis keys auto-expire after 1 hour
- Watcher lock renewal is wall-clock based (30s) not event-count based
- worker_process_init signal warms up vision models on startup
Medium:
- Explicit task_routes for every task name (wildcards never matched)
- app.services.metadata added to Celery include list
- POST /maintenance/recover-stuck endpoint for photos stuck in processing
- Docker healthchecks for worker-light, worker-vision, and Redis
- Task ID in vision log lines for distributed tracing
- Bare except:pass narrowed to specific exceptions
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add active source root directories to the library stats endpoint and
display them in the settings page. Hardcode container PHOTO_DIRS to
/photos since the volume mount handles host path mapping. Add .env to
.gitignore to prevent committing secrets.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Move watch_folders to dedicated 'watcher' queue with its own
single-concurrency container so it never blocks scan/thumbnail slots
- Add get_current_user_media dependency that accepts ?token= query
param for <img src> / <video src> media endpoints (thumb, original,
proxy) — fixes 401 on thumbnails
- Append JWT token to all media URLs in the frontend
- Add missing 'memories' case in sidebar navigation switch
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Centralize execution provider selection in providers.py with
auto-detection and graceful fallback. All ONNX sessions (embedder,
detector, face processor, recognizer) now use the configured providers.
- New VISION_EXECUTION_PROVIDERS env var: "auto" for GPU auto-detect,
or explicit "CUDAExecutionProvider,CPUExecutionProvider"
- Provider priority: CUDA > ROCm > OpenVINO > CPU (when set to "auto")
- docker-compose.yml includes commented-out NVIDIA GPU deploy section
- Supports onnxruntime-gpu as a drop-in replacement for onnxruntime
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Introduce username/password authentication with admin and user roles.
Each user gets their own media directory under /photos/{username}/ with
isolated photos, folders, heaps, and tags. Admins manage users and
observe the full library from a dedicated Settings page.
Backend:
- User model with bcrypt passwords and JWT access/refresh tokens
- Auth router (login, refresh, setup, change-password, status)
- Admin router (user CRUD with last-admin protection)
- user_id FK added to photos, folders, source_roots, heaps, tags
- All data routers scoped by authenticated user
- Scanner inherits user_id from source root owner
- Thumbnails stored under user-prefixed paths for isolation
- Library endpoints accept ?scope=global for admin cross-user view
- Alembic migration 0009 with data migration for existing installs
- Defensive bootstrap.py handles fresh vs existing DB startup
Frontend:
- AuthContext with token lifecycle, auto-refresh, login/logout
- Login page, first-run setup page, auth gate in App.tsx
- Bearer token interceptor on all API requests
- User identity + logout in left sidebar
- Admin-only Settings page with Library Management and Users tabs
- UserManagement panel (add, edit role, reset password, deactivate)
- Settings shows global stats across all users for admin
- Filter bar, right sidebar, keyboard hints hidden on settings page
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Three overlapping fixes so the ingestion pipeline actually runs and the
user can see what it's doing:
Pipeline recovery
- app/database.py: use NullPool when MULITA_CELERY_WORKER=1 so each
Celery task opens a fresh asyncpg connection on its own event loop.
Fixes "another operation in progress" and "Future attached to a
different loop" errors that were dropping ~every thumbnail +
extract_metadata task on the floor.
- app/tasks/thumbs.py: initialize photo=None before the try and rollback
on error so a transport failure in the initial SELECT doesn't raise
UnboundLocalError in the except block and leak rows stuck in 'pending'.
- app/services/vision/bootstrap_models.py: on missing model files,
invoke export_models automatically instead of just warning. First
boot of a fresh install now self-heals.
- app/services/vision/export_models.py: shutil.move instead of
Path.rename so the YOLO export survives the /app → /data/models
cross-volume hop.
- requirements.txt: add ultralytics so export works in a stock image.
Worker topology
- docker-compose.yml: replace the single worker with worker-light
(default/high/low queues, c=2, IO-bound) and worker-vision (vision
queue, c=5, OMP_NUM_THREADS=1 to avoid oversubscription on 6 cores).
Vision is pinned to ≤5 parallel inferences so ONNX doesn't each
spawn an all-cores intra-op pool.
- .env / .env.example: CELERYD_CONCURRENCY replaced with
CELERY_LIGHT_CONCURRENCY + CELERY_VISION_CONCURRENCY.
- Backfill queries in thumbs / scan / vision now ORDER BY taken_at
DESC NULLS LAST so newest photos finish first — the library fills
in top-down in the UI instead of arbitrary insertion order.
Settings visibility
- routers/library.py: new GET /maintenance/pipeline-stats returning
done/total per stage (thumbnails, exif, gps, phash, embeddings,
tags, ocr, faces, face clusters, duplicate groups). Worker-status
now also reports the `vision` queue depth, which was missing.
- services/api.ts: PipelineStats / PipelineStage / ScanStatus types
and the matching client call.
- components/dialogs/SettingsDialog.tsx:
- new Pipeline Progress card with one progress bar per stage
- inline scan banner (processed/total/current folder) inside the
Library section while a scan is running
- Tasks/min throughput computed by diffing worker processed counters
between polls
- Workers section calls out the vision queue and documents the
CELERY_LIGHT/VISION_CONCURRENCY + docker compose up -d scale path
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
PyTorch default install pulls ~7GB of CUDA libs, exceeding disk on small
VMs. Switching to CPU-only saves ~6GB. Also run create_all before alembic
so migrations find existing tables on a fresh Postgres.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Rewrite all vision tasks to use sync psycopg2 sessions instead of
asyncpg — fixes 'another operation in progress' and event loop errors
when Celery forks workers sharing the async connection pool
- Letterbox-pad images to exactly 640x640 for YuNet face detector
(was crashing on non-square thumbnails)
- Deduplicate object detections per label per photo — keep highest
confidence only to avoid photo_tags PK violation on multiple
detections of the same class
- Add all queues (-Q default,high,low,vision) to worker command
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add export_models.py for OpenCLIP ViT-B/32 and YOLOv8n ONNX export
- Fix ArgMax(13) ORT ARM64 incompatibility by passing eot_indices as a
separate ONNX input (computed outside the graph in embed.py)
- Use legacy TorchScript exporter (dynamo=False) for IR version 9 compat
- Upgrade onnxruntime to 1.18.1
- Rewrite bootstrap_models.py with clear separation of auto-downloadable
models (YuNet, SFace) vs manually-exported ones (OpenCLIP, YOLOv8n)
- Wire bootstrap into worker CMD (runs before Celery)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Switch the default database from SQLite to Postgres + pgvector (via
pgvector/pgvector:pg16 Docker image) to support the upcoming vision
pipeline (embeddings, OCR, object detection, face clustering).
- Add `db` service to docker-compose.yml with healthcheck
- Wire `alembic upgrade head` into backend CMD before uvicorn
- Bootstrap empty 0001_baseline revision (schema still owned by create_all)
- Guard SQLite-only PRAGMAs and inline ALTERs behind _is_sqlite flag
- Run `CREATE EXTENSION IF NOT EXISTS vector` on Postgres init
- Add asyncpg, psycopg2-binary, pgvector to requirements
- Provide docker-compose.sqlite.yml escape hatch for legacy SQLite mode
Fresh DB + rescan assumed — no SQLite→Postgres data migration.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The CORS allowed-origins list, host port mappings, log level, container
timezone, and worker concurrency are now all driven by environment
variables with sane defaults. Same-origin access through the nginx
proxy keeps working with no config; direct cross-origin backend
access can be locked down via ALLOWED_ORIGINS.
- backend/config: ALLOWED_ORIGINS env (comma-separated, "*" for any)
exposed via settings.cors_origins. LOG_LEVEL too.
- backend/main: build the CORS middleware from settings.cors_origins,
auto-disable allow_credentials when origins is wildcard (CORS spec
forbids credentials + "*").
- docker-compose: parameterize FRONTEND_PORT, BACKEND_PORT, REDIS_PORT,
CELERYD_CONCURRENCY, LOG_LEVEL, and TZ via ${VAR:-default} so each
has a working fallback if the .env entry is missing.
- .env.example: new template documenting every knob with examples.
- .env: pruned to only the values that diverge from .env.example;
removed dead VITE_API_URL.
- README: configuration knobs table + "accessing from another machine"
section explaining the same-origin proxy story.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Cleans up the maze of overlapping ways folders entered the app, plus
removes the dead trash plumbing left over from the soft-discard
refactor.
Setup model (now)
- ONE env var: PHOTO_DIRS in .env, set to the host path of your
library. Compose mounts that at /photos. That's the entire setup.
- On first boot, the backend auto-creates a SourceRoot row named
"Library" pointing at /photos so the user sees their photos
immediately without configuring anything.
- Source roots and discard live in the database; mulita.yml only
carries operational settings (thumbnails, scanner, performance).
- The "Add Source Folder" dialog is now a directory browser
restricted server-side to /photos and any existing source root —
the user clicks through actual mounted directories instead of
typing container paths they can't possibly know.
Backend
- New services/scanner.bootstrap_default_source_root(): if no
SourceRoot rows exist and /photos is mounted, create one. Wired
into the lifespan handler before cleanup + initial scan.
- New GET /library/browse?path= returning the immediate child
directories of `path`, validated to live under one of the allowed
roots (default mount + every active SourceRoot). Hidden entries
are filtered. Children are tagged with is_existing_root so the UI
can show an "Added" badge. Returns parent path for up-nav, or
null when at the top of the allowed scope.
- scan_all_source_roots now reads from the DB instead of the YAML
config so DB-managed source roots are honoured by initial scan.
- Dropped the placeholder source_roots block from mulita.yml — the
paths /photos/main and /photos/iphone never existed and just
produced startup warnings.
- Dropped TrashSettings, settings.trash, settings.source_roots,
and the SourceRoot pydantic model from config.py. Soft discard
has owned this for a while; it was dead code.
Compose
- Single ${PHOTO_DIRS:-./photos}:/photos:rw mount in both backend
and worker.
- Removed the hardcoded ~/Pictures:/host/Pictures:rw mount — the
PHOTO_DIRS variable is the single source of truth now.
- Removed the trash_data named volume + mounts (no consumers).
- backend/Dockerfile no longer creates /data/trash; it now creates
/data/proxies (which the proxy endpoint actually uses).
Frontend
- AddSourceFolderDialog rewritten as a directory tree picker:
loads /library/browse on open, lets the user navigate up via a
ChevronUp button or down by clicking subfolders, shows the
current path inline, and adds whatever directory is currently
shown. Existing source roots are tagged "Added" so the user
knows what's already registered. Errors from the backend (e.g.
trying to navigate outside the allowed scope) surface inline.
- New library.browse() helper + BrowseChild / BrowseResponse types
in services/api.ts.
Docs
- README Quick Start rewritten around the single PHOTO_DIRS env
var, with macOS/Linux/Windows examples.
- New "How mounted folders and source folders relate" section that
spells out the two-layer model (mount = visibility, source root
= scanning) so the most common confusion is addressed up front.
- Added a "Read-only libraries" subsection that lists exactly which
endpoints fail under :ro.
- "Configuration" section reframed: source roots are managed by the
UI/API now, mulita.yml is operational settings only.
- .env file now has examples for the common host paths.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The first phase-11 file op (inline rename) returns EROFS today
because docker-compose mounts ~/Pictures read-only by default.
Lightroom-style file operations (rename, move, discard-pile empty)
all need to mutate the filesystem, so the right default is :rw.
Flips both the backend and worker mounts to :rw with an inline
comment explaining the trade-off, and adds a "Photo directory
mounts and permissions" section to the README that:
- States the default is now :rw
- Explains exactly which endpoints fail under :ro (rename, empty
discard pile, future move/copy)
- Notes the implication: Mulita has full write access to whatever
host directory ends up at /host/Pictures, same trust model as
Lightroom's catalog folder
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The /photos/{id}/proxy endpoint (added in 1096854) caches transcoded
RAW/HEIC WebPs at /data/proxies/{id}.webp, but the compose file had no
volume mount for that path — files would be lost on every container
restart, forcing repeated full-resolution decodes. Adding a named
volume to both backend and worker so the cache survives restarts.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
