17 Commits

Author	SHA1	Message	Date
Paul Nothaft	8cc7d7d14a	feat(external-media): watch reference folders and import new files automatically (#1345) ... * feat(external-media): watch reference folders and import new files automatically Managed uploads dropped into storage/events/active are picked up by the chokidar watcher; external media had no equivalent, so a NAS folder that keeps growing needed an admin to open the event and press Import every time. Relates to issue 1187. - The import pass moves out of the route into services/externalImportService.js. The watcher and the Import button now run the identical function; the route only validates and maps errors to status codes. - Mutual exclusion is the per-event claim from maintenanceJobState (`external_import:<id>`, seeded on demand by the new ensure()) instead of the in-process Set. The Set stopped a double-click in one process; the claim also stops the watcher on a second replica, or an admin clicking while the watcher is mid-run elsewhere. The run heartbeats so a claim from a dead process is taken over. - services/externalMediaWatcher.js: per-event opt-in via the new events.external_watch column (migration 208), chokidar with awaitWriteFinish so a copy in flight is not imported half-written, debounced full pass per change, a timer sweep every 15 minutes as the fallback for NFS/SMB mounts that deliver no inotify events, optional stat-polling via EXTERNAL_MEDIA_WATCH_POLLING. The set of watched events is re-read every minute, so the toggle works from any replica. A watcher that just started runs one pass immediately. - Deletions are ignored on purpose: a file vanishing from a NAS is at least as likely to be a reorganisation or a dropped mount as an intentional removal, and acting on it would delete a guest-visible photo. Rows whose file is gone stay, as they do today. - Not gated on STORAGE_BACKEND: EXTERNAL_MEDIA_ROOT is always local. - Quiet system passes stay out of the activity log; runs that imported something are logged with actor external-media-watcher. - Frontend: "Watch folder for new files" checkbox under the external folder picker, status line in view mode, EN/DE strings. * fix(external-media): close the review gaps in the folder watcher Codex review of the watcher, round 1. All six findings were real: - Enabling the watcher, or pointing an enabled one at another folder, now requires photos.upload — the permission the manual Import already requires. events.edit alone was a way around it. Only the transition is checked, so a role without photos.upload can still edit an already-watched event. The checkbox is disabled for such roles. - Automatic passes defer files that are still changing: anything modified inside the stability window, or whose size moves across one wait of that window, is left for the next pass. chokidar's awaitWriteFinish only settles the file that fired the event, and the sweep sees no events at all, so a sibling still being copied could be inserted half-written and then skipped forever. - Photos an admin deleted are not brought back by the sweep. The delete routes record the file in external_import_exclusions (migration 209); automatic passes skip the list, the manual Import ignores it and clears it for what it imports. - The six EXTERNAL_MEDIA_WATCH* variables are forwarded in all three compose files; they were documented but the backend services use explicit environment lists, so the kill switch did nothing. - A pass re-checks is_active / is_archived at run time, not only in the minutely reconcile. - The lease is renewed on a timer for the whole run, walk included, and ownership is checked before the event row is touched. * fix(external-media): make automatic passes follow the row, not rewrite it Codex review round 2, four findings, all applied: - The event update route drops non-canonical spellings of external_watch and external_path before the permission guard. SQLite resolves column names case-insensitively, so `External_Watch` reached the column while the guard only looked at the lowercase key. - Exclusions are checked per file at insert time, not against a snapshot taken before the settle wait. A photo deleted during the wait was present in the snapshot and got re-inserted by the loop. - An automatic pass no longer writes source_mode / external_path. It re-reads the row after the walk and the settle wait and stops if the folder changed or the event went managed; the manual Import is the only writer. The options are now `automatic` + `settleMs`. - A pass that deferred files re-arms the debounced import, so a file copied just before the watcher started is not stranded when the sweep is disabled. * fix(external-media): keep exclusions for replaced photos, stop a pass whose event stopped qualifying Codex review round 3, both findings applied: - recordExclusions keys on external_relpath alone. A replaced external photo becomes managed but keeps its relpath on purpose, and deleting that replacement must not republish the NAS original. - An automatic pass checks the full watcher predicate (reference mode, same folder, watch on, active, not archived) before it inserts and on every heartbeat tick during the loop, and stops as soon as the event no longer qualifies. --------- Co-authored-by: Paul Nothaft <paul@MacStudio-von-Paul.local>	2026-09-07 22:29:55 +02:00
Luca	7223118b89	feat(deploy): make the all-in-one image installable without a shell (#1124) ... The all-in-one image could not be installed from a GUI at all — the deployment it exists for. validateEnv treats a missing JWT_SECRET as critical and exits, and the documented run command supplies it with `openssl rand`, a shell command a Synology Container Manager or QNAP Container Station form cannot run. wait-for-db.sh now generates one on first start and persists it next to the database, extending the existing /run/secrets hydration rather than adding a second mechanism. Explicit env still wins, then /run/secrets, then the generated file. The write is load-bearing: JWT_SECRET is exported only when the file actually persisted, because an unpersisted secret would mint a new one every restart and sign every session out. Creation writes to a private temp file and hard-links it into place — atomic, fails with EEXIST when another container won, and the loser adopts the winner's value. Non-regular paths are rejected before the link, since POSIX ln links INTO a directory rather than failing, which would make a mistyped -v target unrecoverable. Also repairs the onboarding paths a new install actually walks: the installer no longer rotates the secrets of a running install on re-run, deprecates the dead scripts/install.sh in place, corrects the CONTRIBUTING dev loop, and fixes the vite proxy target that had been pointing at a stray local port since 0da45e69. Reviewed over three rounds. Co-authored by @Luca-Timo.	2026-08-22 18:37:12 +02:00
Paul Nothaft	b69dd134d0	feat(faces): People in this gallery — face recognition via an optional ML sidecar (#1074) (#1075) ... * feat(ml): optional face-detection sidecar, opt-in and inert by default (#1074) First of four PRs for "People in this gallery". This one ships only the sidecar, its wiring and its CI — no schema, no backend code, no UI. Nothing in PicPeak calls it yet. picpeak-ml is a single FastAPI + onnxruntime container: three endpoints (/health, /info, /faces), no database, no volumes, no egress, no model download at runtime. Clustering, person identity and every privacy decision stay in the backend where the data already lives. Models are YuNet (detection) + FaceNet-512 (embedding), both MIT, both pinned by URL and SHA-256 and verified at build time. The licence analysis is in ml/LICENSES.md: the more accurate InsightFace weights are non-commercial-only and PicPeak's users are working photographers, so they are never baked into an image we publish. Two things worth review attention: - Alignment uses a least-squares similarity transform (Umeyama), NOT cv2.estimateAffinePartial2D. RANSAC and LMEDS exist to reject outliers among many correspondences; given five landmarks and no outliers they fit a three-point subset exactly and let the rest drift. Measured on a real off-frontal portrait: eyes and nose pinned to 0.11px, mouth corners 11.8px out on a 160px crop. Umeyama distributes it (max 6.5px, rms 5.1 vs 7.4). The failure mode is silent — a bad warp still yields 512 confident floats — so tests/test_pipeline.py pins it numerically. - FACENET_ONNX_URL has no default and the build fails loudly without it. deepface distributes FaceNet-512 as Keras .h5 only, so the ONNX is produced once by tools/convert_facenet.py and published as a release asset. Converting inside the build would drag TensorFlow through both architecture legs of every build to produce a byte-identical file. The CI jobs are gated on the FACENET_ONNX_URL repository variable and skip cleanly until it is set. Off by default, twice over: the sidecar is behind the `faces` compose profile, and the backend will gate on a `faces` feature flag that defaults to false. FACE_ML_URL defaults to http://picpeak-ml:8000 so the standard deployment needs no configuration — nothing dials that host while the flag is off, which is why a non-resolving default is harmless. Verified: 27 pytest tests green; YuNet loads and detects against a real portrait with its landmark order matching the alignment template index-for-index; both compose files validate and the faces profile is correctly excluded from a default `up`; workflow YAML parses and the job graph resolves. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * fix(ml): pin the converter toolchain, verify parity, drop a false reproducibility claim (#1074) Ran the FaceNet-512 conversion for real and corrected what the previous commit assumed about it. The conversion works: 23,497,424 parameters, 89.6 MB ONNX, and the converted graph matches the Keras original to 2.086e-06 absolute / cosine 1.0000000000. That check is now part of the script rather than something I did once by hand — a subtly wrong graph still returns 512 plausible floats, so it refuses to leave the file on disk if parity fails. Also ran the full pipeline against both real models end to end. The embedding is L2-normalized to 1.000000, and the same face survives being re-rendered: half scale 0.973, double scale 0.984, JPEG q40 0.987, rotated 8 degrees 0.984, brightness +40 0.988. Scale invariance in particular is evidence the alignment warp is doing its job. Corrected claim: the conversion is NOT byte-reproducible. Two runs with the same pinned versions on the same machine gave different SHA-256s. The graphs are functionally identical — same 336 nodes, same 271 initializers, every weight matching to 0.000e+00 — but a few initializer names differ because tf2onnx's traced-op naming is not deterministic (Keras layer naming is deterministic; I checked). The previous commit message and README both claimed byte-identical output. They were wrong, and it matters: anyone re-running the conversion gets a different hash, and without this note that reads like tampering. The build-time SHA-256 pins one published artifact so its URL cannot start serving different bytes; validating a fresh conversion is the parity check's job. requirements-convert.txt now pins the exact set that produced the artifact, including transitive keras/protobuf/numpy, and documents that the converter needs Python 3.11 while the image runs 3.12. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * feat(faces): schema, queue, clustering and API for People in this gallery (#1074) Backend half of the feature. Migration 177, a face-detection queue, the clustering engine, the gallery and admin APIs, and the privacy wiring. No UI yet; nothing is reachable until the `faces` feature flag is on, which defaults to false. The flag is the gate, not FACE_ML_URL. That variable now has a working default (the compose service name), so its presence proves nothing about intent — if it were the gate, every install would poll a hostname that does not resolve. faceQueue re-checks the flag every tick, so turning it off stops the workers without a restart. Visibility scoping is the part worth reviewing closely. Face rows have no concept of photo visibility, but guests are restricted to photos.visibility='visible'. A raw count leaks how many hidden photos someone appears in, and an unscoped cover face renders a crop of a photo the guest may not open — with the best-scoring face being the likeliest pick, so it would happen often rather than rarely. facePeopleService recomputes both per request against the caller's own scope, and event_people.face_count_total is named to be conspicuous in a guest path. Six tests cover it, including the case where a person's photos are ALL hidden and they must vanish entirely. Face data is excluded from backups and .picpeak exports, per the decision in the thread: it is derived, so a restore re-scans rather than carrying biometrics between operators. Three separate mechanisms, because the engines cannot be filtered alike — EXCLUDED_TABLES for export, --exclude-table-data (not --exclude-table; the CREATE TABLE must survive or restore breaks on the first query) for Postgres, and DELETE + VACUUM on the temp copy for SQLite, which has no way to exclude a table from a whole-file .backup. The VACUUM is not cosmetic: without it the pages stay in the file and the claim is false on disk. Archiving now purges face data explicitly. photo_faces cascades off photos, but archive deletes neither the photo rows nor the event, so without this an archived gallery kept its biometrics indefinitely. Other decisions: clustering keeps names across a re-cluster by majority inheritance (without it, one button click silently discards every name the photographer typed); consolidation refuses to merge two people who were named differently; assignment never compares across model_version, since embeddings from two pipelines are not comparable; low-quality faces are stored but left unassigned so they show in "this photo contains" without spawning junk people. Migration is 177, not 174 — 174/175/176 landed on main while this branch was open. 29 tests green: 7 migration (idempotency, down(), cascade, and that installing it enqueues NOTHING), 11 clustering, 11 privacy/visibility. Lint clean; the pre-existing error counts in databaseBackup.js and server.js are unchanged. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * feat(faces): People strip, face filter and admin controls (#1074) Frontend half. Renders nothing anywhere unless the `faces` feature flag is on AND the photographer enabled detection for the gallery — the whole guest surface hangs off one boolean, event.people_enabled, which the server computes from the flag, the per-event toggle and the show-to-guests toggle together. Guest side: a People strip between the filter bar and the grid, circular crops from each person's cover face, an active-filter chip row, and a "Show all" bottom sheet. The face filter composes with category, search, media type and the liked/saved/rated filters in the same useMemo rather than replacing them, so "photos of Anna that I liked" works. Two people selected means AND by default — that is what picking a second face almost always asks for — with a toggle to OR that appears only once a second person is picked. Unnamed people show a photo count and never "Person 7". A number is honest about what the system knows; an invented name is not. There is a test asserting we don't do it. The strip renders nothing below two people, collapses to one line when dismissed (persisted per slug, so dismissing one gallery says nothing about the next), and appears mid-backfill with a progress line rather than blocking the gallery behind a spinner. Avatar crops are computed in ratios of the source dimensions so they survive whatever rendition the browser gets; without width/height they fall back to an uncropped thumbnail, since a wrongly-offset crop is worse than no crop. No new download endpoint: "download these N" rides the existing photoIds path, which already enforces access level and per-category permissions server-side. Adding a person_id selector would have been a second thing to authorize for no gain. Guest-facing copy never says "biometric" or "recognition" — those words describe our implementation, not the guest's experience. The sheet's footnote answers the first question every guest has (where does this go?) inline. The admin card, by contrast, is explicit: it states the controller obligation next to the toggle, and warns that scanning materializes the preview tier on galleries that never generated one, which is real CPU and disk an admin should know about before a 2,000-photo backfill. EN + DE translations. 140 frontend tests green (8 new), tsc and eslint clean. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * fix(faces): measured match threshold, working build defaults, 89MB smaller image (#1074) Ran the Phase 0 spike that had been outstanding, published the model, and fixed what both turned up. THRESHOLD IS NOW MEASURED, NOT GUESSED. LFW's standard 1000-pair protocol run through this exact pipeline (YuNet -> Umeyama alignment -> FaceNet-512 ONNX), 100% detection on 2000 images: same person cosine 0.6958 +/- 0.1415 diff person cosine 0.0849 +/- 0.1674 separation 0.6109 peak accuracy 96.60% @ 0.405 So the pipeline separates people well — the thing I could not previously claim, since every earlier number was the same face re-rendered. Default moves 0.62 -> 0.50. The old value was a placeholder and a bad one: it gave 0% false merges but 22.4% false splits, i.e. roughly one in four same-person pairs failing to join, which fragments a gallery badly. 0.50 gives 1.0% false merge / 8.2% false split. Peak accuracy (0.405) is deliberately NOT chosen: for clustering the two errors do not cost the same. A false split is a duplicate row the photographer can merge away; a false merge puts a stranger into someone's "download my photos" — and until the Phase 2 merge/split UI ships, there is no way to undo one. So this sits on the conservative side of the optimum. The spike is committed as ml/tools/benchmark_threshold.py rather than thrown away, so "why 0.50?" has an answer in six months and a re-tune is one command. BUILD DEFAULTS. FACENET_ONNX_URL/_SHA256 now default to the published ml-models-v1 release asset, so `docker build ml/` and `docker compose --profile faces up` work with no arguments. Blanking either still fails loudly — a URL without a checksum is never acceptable, since the checksum is what makes the URL safe to trust. Found by running compose for real: it failed exactly as designed, which was correct behaviour and a bad out-of-box experience now that a canonical artifact exists. IMAGE SIZE. 389MB -> 300MB single-arch. `chown -R` after COPY rewrote every copied file into a fresh layer, duplicating the 90MB model for nothing; the user is now created before the copies and ownership set via COPY --chown. Also drops pip/setuptools from the runtime image. Measured RSS is 186MiB idle, and the container answers /faces end-to-end in well under the 80-150ms/photo the issue budgeted. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * fix(faces): threshold 0.50 -> 0.60 from real clustering, theme-aware People strip (#1074) Both fixes come from running the feature on an actual gallery — 61 photos, 5 real identities — rather than reasoning about it. THRESHOLD. The LFW pairwise sweep in the previous commit said 0.50, and it was wrong. On a real gallery at 0.50, three of six visible clusters were contaminated: two different people merged into one strip entry, which is the exact failure that puts a stranger into someone's "download my photos". Pairwise error rates do not predict cluster purity. Greedy assignment compounds — one wrong face drags the centroid toward the midpoint between two identities, making the next wrong face likelier. A 1% pairwise false-merge rate is not a 1% chance of a clean gallery, and no amount of staring at an ROC curve would have shown that. Sweep against ground truth (5 identities): 0.50 -> 6 clusters, 3 contaminated 0.56 -> 6 clusters, 0 contaminated 0.60 -> 5 clusters, 0 contaminated <- exactly right 0.64 -> 5 clusters, 0 contaminated, fewer faces assigned 0.60 recovers the right number of people with no contamination; higher only loses coverage. Migration 177 carries the full reasoning so the next person to touch this knows why the obvious pairwise answer is the wrong one. THEME. The People strip hardcoded `text-neutral-800` for named people. On a dark gallery — which the screenshot immediately showed — that renders a named person's label almost invisibly, while UNNAMED people stayed legible. Exactly backwards. Labels, headings, the collapsed summary, the scan line and the filter chip row now read the gallery's own theme tokens (--color-text / --color-muted-text / --color-accent / --color-surface-border) like the rest of the gallery surface. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * fix(faces): keep the mobile filter row inside the viewport (#1074) At 390px the photo count and Clear link were pushed against the right edge by ml-auto and clipped. Only apply it from the sm breakpoint up, where there is room; below that they flow after the chips. Found by screenshotting the real thing on an iPhone-sized viewport. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * feat(faces): complete Phase 1, add People management and auto-categories (#1074) Closes the two Phase 1 gaps, then builds Phase 2 and Phase 3. PHASE 1 GAPS. "Download these N" was specified, described as done in an earlier summary, and never actually built — I had verified the backend needed no new endpoint and let that stand as if the button existed. It now hands the filtered photo ids to the same path as a manual selection, so the server re-applies access level and per-category permissions on the way through. Photos in a downloads-disabled category are excluded client-side too, so the number on the button is the number the guest receives. Hidden entirely when downloads are off for the gallery. Lightbox person chips ("In this photo: Anna") are the second way into the face filter — a guest looking at a photo of themselves can act on it without scrolling back to the strip. Tapping one closes the lightbox and filters the grid behind it. PHASE 2. A People management modal over the endpoints that already existed and were already tested: rename inline, merge (multi-select, first pick is the target so the name a photographer typed survives), split via a face picker, hide, ignore. This matters more than it sounds — clustering deliberately errs toward splitting because a wrong merge puts a stranger into someone's download, and that trade only works if merging is easy. PHASE 3. Rule engine over face_count plus face-area ratio: 0 -> Details, 1 large -> Portraits, 2-5 -> Small groups, >5 -> Groups. The area ratio is what separates "a portrait of someone" from "someone is in this landscape". Three guarantees, all tested: it only ever fills an EMPTY category (enforced in the query AND re-checked in the UPDATE, so a photographer setting one mid-run still wins), everything it touches is marked auto_categorized so undo is exact, and it is a no-op unless separately enabled. Migration 178 adds the column — separate from 177, which has already run wherever this branch is deployed. Verified on the real gallery: 61 photos -> 48 portraits + 13 small groups, undo cleared exactly 61 and left the manual ones alone. Merge moved faces and removed the source. Both confirmed against the database, not just the UI. TWO BUGS THE BROWSER CAUGHT, both invisible to tsc: - The lightbox destructure never landed — my patch targeted a line that has a default value, matched nothing, and failed silently. `people` resolved to something else entirely and the chips would never have rendered. eslint's "outer scope value" warning is what surfaced it. - Admin face thumbnails 403'd because <AuthenticatedImage> attaches whatever gallery token is in session storage; an admin who has also opened one of their own galleries sends a type:"gallery" bearer to an admin route. Admin routes authenticate from the httpOnly cookie, which a plain same-origin <img> sends by itself. Worth noting AdminPhotoGrid has the same latent shape; not touched here. Also: the admin card now reports "N people (M shown to guests)" when those differ, so the settings page and the gallery stop disagreeing without explanation. 45 backend tests (8 new) and 140 frontend tests green; tsc and eslint clean. EN + DE for every new string. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * perf(faces): batch migration DDL and drop the face stack from server.js import (#1074) CI's backend job timed out at 10 minutes on the first run of this branch. Nothing failed — 132 of 182 suites passed and the wall clock ran out. Main does the same 182 in 124s, and where main has 12 suites slow enough for jest to print a duration, this branch had 77. Two changes, both worth making regardless of how much of the gap they close: - Migration 177 added its columns one ALTER TABLE at a time (four on photos, three on events, plus a separate index statement) and seeded settings with a SELECT and an INSERT per key. It now uses one alterTable per table and one SELECT plus one bulk INSERT. 178 folds its index into the same statement as its column. That chain replays in ~90 suites, so statement count there is multiplied by 90. - server.js required faceQueue at module scope, which pulls in axios and — through imageProcessor — sharp. Every supertest suite that imports server.js was paying for a module graph it never uses. Now required inside the startup block, next to the call that needs it. Honest about the evidence: locally the migration delta measures at zero (1.15s vs 1.13s for the same suite, three runs each), so batching alone does not explain an eight-minute regression. A fast local disk and many cores mask per-statement and per-import costs that a two-core runner with a shared disk does not. These are the two real costs this branch added to a path that runs in almost every suite; whether they are sufficient is a question for CI, not for another round of local speculation. 37 face tests still green after the change. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * i18n(faces): complete EN and DE coverage for the face feature (#1074) The admin card and the Features toggle were rendering entirely from inline English `defaultValue` fallbacks — 22 keys existed in no locale file at all, so a German admin saw an English consent notice, English toggles and English buttons. The gallery side was already translated; the admin side was not, and nothing in the toolchain flags this because a `defaultValue` always renders something. Adds the missing `admin.faces.*` (19), `settings.features.faces.*` (2) and shared `common.clear/saved/saveFailed` in both languages. Existing keys are left alone (setdefault, not overwrite), so the shared `common` strings other features rely on are untouched. Committed the audit as frontend/scripts/i18n-faces-audit.py rather than throwing it away: it extracts every t() key the face components actually use and diffs it against each locale, and it also reports German values that are byte-identical to English, which is the usual shape of an untranslated copy-paste. Currently: 69 keys in use, EN complete, DE complete, no identical pairs. Verified in the browser, not just in the JSON — the German card reads "61 / 61 Fotos durchsucht · 16 Personen (5 für Gäste sichtbar)" end to end. Also checked the components for hardcoded user-facing text (JSX nodes, title/aria-label/placeholder attributes) outside t(); there is none. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * fix(faces): 13 defects from external review — coordinates, counts, erasure, races (#1074) Codex reviewed the branch against main. Thirteen findings, nine P1. I checked every one against the code and could not dismiss a single one as a false positive, so all thirteen are fixed here. THE WORST ONE: bounding boxes were stored in the wrong coordinate system. The sidecar reports coordinates in the space of the image it was HANDED — which is the ≤1920px preview, not the original — while every consumer compares them against photos.width/height, the original dimensions. A 6000px photo therefore produced boxes ~3x too small and areas ~9x too small: avatar crops landed in the wrong place and the Portraits rule could never fire. It is invisible on any photo already under 1920px, which is exactly why the demo gallery and every screenshot looked correct. Now scaled once in faceProcessor so everything downstream can assume original-image coordinates. ERASURE. The FK cascade on photo_faces is decorative on SQLite: PicPeak never enables `PRAGMA foreign_keys`, so deleting a photo left its embeddings behind. I first enabled the pragma globally and reverted it — six unrelated suites immediately failed on pre-existing dangling references, and switching it on would start rejecting inserts on every existing install. That is a real change worth making, but it is its own PR, not a rider on this one. Instead deletion purges explicitly: purgePhotoFaces in the photo paths (single, bulk, service) and photo_faces/event_people in deleteEventCascade. Tests assert this with the pragma explicitly OFF, so they can only pass if the code does the work. COUNTS. A re-scan deleted the old face rows without undoing their contribution to event_people, so counts inflated on every re-scan and ghost people survived. Now the affected people are recomputed before the replacements are assigned. My own "must not double its faces" test only checked photo_faces rows, which is why it passed throughout. RACES. A worker that finished after an admin purged the event committed its rows anyway — erasure reported success and the data reappeared. The commit is now conditional on the row still being 'processing'. And assignFaces is read-modify-write over an event's people, so two workers lost each other's updates; it is now serialised per event with an in-process mutex plus a Postgres advisory lock for the multi-pod case the queue advertises. METADATA LOSS. Merging discarded the source's name and suppression flags, so a merge could erase a typed name or un-hide someone. Reclustering remembered only people with a label, so an unnamed-but-hidden bystander came back guest-visible after one "Re-group people" — and suppression now propagates to every descendant cluster, not just the majority one. Also: export reset face_status so a restored gallery re-scans instead of claiming to be scanned forever; manual category edits clear auto_categorized so "undo automatic" cannot delete a photographer's own choice; external photos are skipped rather than failed (resolvePhotoStorageKey returns null for them by design); the gallery refetches photo memberships as a scan progresses so filtering is not stale; a failed VACUUM now fails the backup rather than publishing one that may retain biometric pages; and the ML Dockerfile's `|| true` is scoped to the uninstall — as written it was `(install && uninstall) || true`, so a failed dependency install produced a green layer and an image with no onnxruntime. Four new regression tests. Full backend suite failure set verified identical to origin/main; frontend 140 green; tsc and eslint clean. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * fix(faces): 12 more defects from review round 2 — cross-event purge, leaks, lifecycle (#1074) Second Codex round on the same diff, now including round 1's fixes. Twelve findings, seven P1. Again none were false positives. SECURITY, AND MINE FROM ROUND 1: the bulk-delete face purge iterated the raw `photoIds` from the request instead of the event-scoped `photos` rows the handler had already validated. purgePhotoFaces has no event scope of its own, so an editor could pass another gallery's photo id and delete its face data — even though the photo deletion right below it was correctly scoped. Fixing one thing and introducing another is exactly why the second round was worth running. ANOTHER VISIBILITY LEAK, same class as the one round 1 fixed: /people returns scan progress, and getScanStatus counted every photo with a face_status — including hidden ones. Guests could read the hidden-photo count off the progress bar while the people list and covers beside it were properly scoped. Now scoped by the same predicate, with the caller passing its audience. RECLUSTER, ROUND 1'S FIX WAS INCOMPLETE. I made suppression follow every descendant but still copied the flags from the majority ANCESTOR. When reclustering merges a visible named person with a hidden one, the majority ancestor is often the visible one — republishing the hidden person's photos. Suppression is now OR-ed across every ancestor contributing faces. The name also now goes to the genuine largest descendant; the previous code took whichever cluster came first in map order, which the comment already claimed it did not. LIFECYCLE. Face data is excluded from backups and exports, but photos. face_status came across intact, so a restored install claimed every photo was scanned while holding no faces — and the worker only claims 'pending', so it stayed that way forever. Now: the SQLite backup requeues in the dump, restore requeues after the pool reinit (the Postgres path cannot rewrite rows inside pg_dump), the portable importer purges LOCAL face tables (they were excluded from the replace list, so another instance's embeddings survived an import with FK checks suspended) and requeues, and archiving disables detection so a restored archive is honestly off rather than enabled-and-empty. WRITE PATHS. Only processPhoto enqueued. The synchronous upload path (chunked-upload completion, watch-folder) left photos unscanned, and replacePhoto kept the OLD image's faces on a row now pointing at a different picture — stale identities shown on the new photo. FRONTEND. PeopleSheet and the admin manager rendered centred thumbnails and ignored the bbox, so on group photos the avatar showed whoever stood in the middle and two people from one photo were indistinguishable — in the manager whose entire job is telling faces apart. The crop maths is now one shared helper (faceCrop.ts) so the three surfaces cannot drift again. Full-page layouts (gallery-premium, gallery-story) render their own lightbox and never received the people props. Backend failure set verified identical to origin/main; frontend 140 green; tsc and eslint clean. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * fix(faces): round 3 — five of round 2's fixes were wrong or no-ops (#1074) Third and final Codex round. Eight findings, four P1 — and the important part is that FIVE of them are defects in round 2's fixes, not in the original code. - The sync-upload enqueue I added was a silent no-op. It queried through `trx` after the transaction had already been committed, which throws "Transaction query already complete" straight into the catch I had wrapped it in. Chunked uploads and watch-folder imports were still never scanned, and the code read as though they were. Uses `db` now. - The post-restore requeue ran BEFORE the files were restored, in both the portable importer and the native restore. The face worker is live during a restore, so it could claim those rows and scan the previous instance's files, or fail them for originals not yet on disk — with nothing to requeue them afterwards. Both now run after file restoration; the native one is extracted into requeueFaceScans() and called from the full and database-only paths. - The admin face crop mixed coordinate spaces: an original-pixel bbox scaled against the THUMBNAIL's natural size. The API now returns the source dimensions alongside the box, so there is one space to reason about. - Forwarding people props through layoutProps did not make them work — the full-page layouts never destructured them. GalleryStoryLayout now threads them to its own lightbox. Genuinely new findings, all in the same class as ones already fixed: - releaseToPending updated unconditionally, so a photo purged while its sidecar request was in flight came back as 'pending' and was rescanned — biometric rows reappearing after the purge reported success. Round 2 fixed exactly this on the COMMIT path and I did not carry it to the retry path. Now guarded on 'processing'. - purgePhotoFaces left face_status alone, so a worker mid-scan still satisfied its commit guard and could write fresh faces into a photo being deleted — orphans, since the FK cascade is inert on SQLite. It now clears the claim as part of the purge. - Phase 3 was unreachable: the migration seeds face_auto_categorize_enabled false and nothing could ever write it, so the rule engine and its undo endpoint returned "disabled" in every real flow. Added GET/PUT and a toggle on the admin card, EN + DE. NOT fixed, deliberately: GalleryPremiumLayout uses yet-another-react-lightbox rather than the shared PhotoLightbox, so person chips there are a real port rather than a prop forward. Recorded as open rather than bodged. Backend failure set identical to origin/main; 41 face tests and 140 frontend tests green; i18n audit reports EN and DE complete at 71 keys. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * feat(faces): block face recognition on the all-in-one image (#1074, #1042) The single-container image cannot run this feature, so it is refused there rather than left to degrade. WHY, since the reason is not obvious from the code: the AIO image runs the backend, the frontend, SQLite and every background worker inside one container aimed at "one photographer plus guests browsing". It has no Redis, SQLite gives it a single writer, and it contains no ML sidecar to talk to. Face detection would add a second image-processing pipeline competing with Sharp for the same CPU and memory. That failure is not loud — the install just becomes slow and looks broken, which is the worst possible shape for a deployment whose whole promise is one container and no decisions. Gated on an explicit PICPEAK_SINGLE_CONTAINER marker, NOT inferred from SERVE_FRONTEND or a SQLite path: plenty of legitimate multi-container setups serve the frontend from the backend or run SQLite, and none of them should lose the feature by accident. Three layers, because the first is the only one that enforces: - faceSettings.isFeatureEnabled() returns false before consulting the flag, so a database restored from a full deployment with `faces` enabled still cannot switch it on here. - The feature-flag API forces `faces: false` in both directions, so the admin UI reflects reality instead of offering a switch that refuses to stay on. - The Features tab renders the card disabled with a plain-language reason, read from a new `single_container` field on /admin/system/version (an endpoint the admin UI already calls). Documented in ml/README.md and .env.example. Three tests pin the behaviour, including that the marker only accepts explicit truthy values. NOTE FOR PR #1068: this expects `Dockerfile.aio` to set `ENV PICPEAK_SINGLE_CONTAINER=true`. That one line lives on that branch and is not in this commit — until it lands, an AIO build would still offer the feature. Worth adding alongside the `Limits` section of docs/single-container.md. 44 face tests green; EN + DE complete at 72 keys. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc * test(faces): pin the bbox coordinate space with a real scale factor (#1074) The coordinate-space bug — boxes stored in preview space while every consumer reads them as original-image pixels — had no test, and could not have been caught by the ones that existed: every photo in the demo gallery is 750px, so the scale factor was always exactly 1.0 and the correction never executed. Verified by hand first, on a real 4000x3000 upload with the face placed off-centre so a wrong crop would be unmistakable. Before the fix the stored box was 1493,204 (preview space, face actually at x≈2850-3618); after, 3110,426 — a factor of 2.083, exactly 4000/1920, landing inside the face. The admin crop then resolved to left=-395px/top=-46px on a 64px window, which is the face centred. That verification is now a test rather than a memory. Three cases: a 4000px photo must scale by 4000/1920, a 1920px photo must NOT change (the case that hid the bug), and a row with no width must fall back to unscaled rather than storing NaN. Note for anyone extending these: jest hoists mock factories above the file, so anything they close over has to be `mock`-prefixed. Getting that wrong fails at transform time with a message that does not name the variable. 47 face tests green. Claude-Session: https://claude.ai/code/session_01Ra4hcsYiKuQLbbRsg6EjAc --------- Co-authored-by: Paul Nothaft <paul@MacStudio-von-Paul.local>	2026-08-18 22:37:28 +02:00
Paul Nothaft	8337a716b1	fix(file-watcher): bound concurrent photo processing (#846) ... * fix(file-watcher): bound concurrent photo processing chokidar fires 'add' once per file — with no ignoreInitial option the boot scan fires it for every existing file, and a bulk drop into the watch folder fires it for every new one at once. Each handler runs DB lookups plus (for new files) a full sharp pipeline; sharp.concurrency(2) only caps libvips threads WITHIN one operation, not the number of parallel pipelines, so unbounded handlers can OOM small hosts. Gate both 'add' and 'unlink' through a shared p-limit (FILE_WATCHER_CONCURRENCY, default 2, floor 1) — mass deletes otherwise burst DB work and ZIP-cache invalidation the same way. p-limit is pinned to ^3.1.0, the last CommonJS release. Adapted from the filpgame fork (426ca491) — thanks @filpgame; extended to cover 'unlink', documented in .env.example, plus a lock-in test for the existing Sharp cache/concurrency caps this bound relies on. * chore(compose): pass FILE_WATCHER_CONCURRENCY into the backend container (codex review of #846) The backend service uses an explicit environment list (no env_file), so the documented override never reached the container in the default compose deployments. Added to both compose files + root .env.example.	2026-07-19 22:00:46 +02:00
Luca	286975dc52	fix(setup): address PR #714 review — password UX, script token, race, nits ... Blockers: - SetupPage now mirrors the server password rule (>=8 with upper/lower/digit) so a green client isn't bounced by the server; server errors carry a `field` (routes/setup.js) that the client maps to a translated key instead of rendering raw English. New i18n: setup.invalidToken, setup.passwordRequirements. - picpeak-setup.sh: the ADMIN_CREDENTIALS.txt block no longer dead-ends on the wizard path — when no legacy admin was seeded it prints the one-time setup token (from data/SETUP_TOKEN / docker compose logs) and points at /setup. Concern: - createInitialAdmin creates the admin + burns the token in ONE transaction, atomically claiming the token (null-if-present, expect 1 row) so a double-submit can't create two super_admins. Cross-DB (whereNotNull, trx-only writes). Added a concurrency test. Nits: - SetupPage redirects to /login when /setup/status errors (no form flash on a configured instance). - Dropped the unused DATABASE_URL from docker-compose.yml. - Documented why secrets are chmod 644 (three different reader users).	2026-07-02 13:21:55 +02:00
Luca	415bffa04c	feat: zero-config first run — in-browser admin bootstrap + auto-generated secrets ... Fresh installs need nothing in .env. See PR description for the full feature.	2026-07-01 14:49:18 +02:00
Paul Nothaft	0205c7dcce	chore: migrate Docker registry + GitHub URLs to PicPeak org ... Repo transferred from the-luap/picpeak → PicPeak/picpeak. Docker images publish to ghcr.io/picpeak/picpeak/{backend,frontend} (lowercase, per the GHCR canonical form computed by docker-build.yml's `${GITHUB_REPOSITORY,,}`). Sweep covers: - docker-compose.production.yml + Dockerfiles → new image registry path - README, CONTRIBUTING, SECURITY, SIMPLE_SETUP, scripts/picpeak-setup.sh → new GitHub URLs - Update-check / release-notes services (updateCheckService, environmentService, updateNotificationService, adminSystem, UpdateNotification, githubReleaseUrl) → GitHub API + tag URLs use the canonical PicPeak/picpeak path - Issue templates + README-DOCKER + workflow README → updated package URLs - One commit-context comment in migrations/090 + customerAccountsService CHANGELOG.md is intentionally untouched (historical release entries are immutable; GitHub auto-redirects the old URLs indefinitely). CLAUDE.md keeps the bare `(the-luap)` reference — that's the maintainer's personal handle, not a repo URL. 22 files, 48/48 line swaps (every change is a 1:1 URL replacement).	2026-06-29 20:20:13 +02:00
Paul Nothaft	714a9f6fb1	fix(upload): auto-throttle on low-memory hosts + correct documented RAM minimum (#628) ... The README claimed 2GB RAM as the minimum, but two background-processor worker loops × sharp.concurrency(2) means up to four libvips threads can decode full-resolution images in parallel — peak RSS lands at 1.5GB+ on a batch of 20MP+ photos. Add Postgres + Redis + Node baseline and one heavy batch on a 2GB VPS OOM-kills the backend, surfacing as 503s on thumbnails until restart:unless-stopped brings it back. Reported in #602, filed as #628. Three changes, smallest-surface-area each: 1. backgroundProcessor.js — on startup, when UPLOAD_PROCESSOR_CONCURRENCY is NOT set and os.totalmem() reports < 3GB, default to 1 instead of 2 and log a one-shot warning naming the override env var. Explicit env-var setters keep their value. os.totalmem() reports container memory under cgroup v2 so this works in Docker / k8s as well as bare metal. 2. README.md — bumped the documented minimum from 2GB to 4GB, kept 2GB only as a "Low-memory hosts" recipe pointing at UPLOAD_PROCESSOR_CONCURRENCY=1 with the throughput trade-off spelled out. Added the 503-on-OOM symptom so the next reporter finds it via search. 3. docker-compose.production.yml — commented mem_limit / memswap_limit example on the backend service. Off by default (don't surprise existing deployments) but visible to operators thinking about shared/multi-tenant hosts. restart:unless-stopped already on every service. No code path for memory-aware runtime throttling (Luca's option 4) — out of scope for a bug fix; tracked separately if #1-#3 don't close the case.	2026-06-17 23:04:30 +02:00
Paul Nothaft	efa6b4a205	fix(brand-title): runtime substitution so GHCR-image users can override (#521 follow-up) ... @Rekoo-PS confirmed the prior #521 fix landed on beta but reported the preview still shows the default "PicPeak" title — their brand is "arkan-studio". Root cause: that fix used Vite's build-time %VITE_DEFAULT_TITLE% substitution. Self-hosters running the pre-built ghcr.io/the-luap/picpeak/frontend image can't override at build time without rebuilding, so they were stuck with whatever the upstream build baked in. Pivot to runtime substitution: the frontend container now reads BRAND_TITLE / BRAND_DESCRIPTION env vars on startup and envsubsts them into index.html. Change the values in .env, restart the frontend service, done — no rebuild required. Mechanics: - frontend/index.html: tokens are now ${BRAND_TITLE} / ${BRAND_DESCRIPTION} (shell expansion syntax, passes through Vite unchanged into the built dist). - frontend/Dockerfile: install gettext (provides envsubst), snapshot /usr/share/nginx/html/index.html → index.html.tpl at build, install docker-entrypoint.sh, wire ENTRYPOINT to it. The .tpl is the immutable source — every container start re-renders index.html from .tpl, so restarts pick up new env values cleanly (no accidental "first-boot env stuck forever" trap). - frontend/docker-entrypoint.sh: applies defaults if env unset, runs envsubst (locked to BRAND_TITLE + BRAND_DESCRIPTION explicitly so /assets/*.js template literals aren't touched if anyone ever extends substitution to the bundle), execs nginx. - frontend/vite.config.ts: drop the htmlTitleDefaults plugin — no longer needed since substitution is fully runtime. - frontend/.env.example + .env.production.example: drop the VITE_DEFAULT_* docs (the vars no longer have effect). - docker-compose.yml + docker-compose.production.yml: pass BRAND_TITLE / BRAND_DESCRIPTION env into the frontend service with sensible defaults so unconfigured installs work unchanged. - .env.example: add BRAND_TITLE / BRAND_DESCRIPTION with comment pointing at the social-preview use case. Verified end-to-end against the built image: - BRAND_TITLE="Arkan Studio" BRAND_DESCRIPTION="Wedding photographs by Arkan Studio" → index.html serves <title>Arkan Studio</title> + og:title="Arkan Studio" + og:description correctly substituted. - .tpl preserves ${...} tokens so the next restart can re-substitute. - Bundle assets unaffected. - Defaults applied when env unset → <title>PicPeak</title>. Docs PR in picpeak-docs describes the two new env vars under "Social link preview fallback" in the environment-variables reference. Refs: #521	2026-05-21 10:00:21 +02:00
Paul Nothaft	0b0b1bb2d5	fix(install): silence pg healthcheck noise + drop legacy workers container (#484) ... Three install-experience bugs that compounded into MrGabri's "fresh install fails" report: 1. **postgres healthcheck noise.** `pg_isready -U <user>` without -d defaults to probing a database whose name matches the user. Since DB_NAME defaults to picpeak_prod (not picpeak), every healthcheck interval logged FATAL: database "picpeak" does not exist into postgres logs even though the install was working correctly. Reporter saw the FATAL, assumed broken, restarted with DB_NAME=picpeak, hit a tainted-state migration error on the second try, filed a bug. Fixed in both docker-compose.production.yml and the inline compose generated by scripts/picpeak-setup.sh — pin -d to ${DB_NAME} so the probe hits the real database. 2. **backend container shows perpetually `unhealthy`.** Both compose files used `curl -f` for the backend healthcheck, but backend/Dockerfile only installs dumb-init + postgresql-client + ffmpeg — no curl. Switch to wget --no-verbose --tries=1 --spider to match what backend/Dockerfile's own HEALTHCHECK already does. Now docker ps, docker compose ps, and the backend image's built-in healthcheck all agree. 3. **stale separate `workers` container.** scripts/picpeak-setup.sh still generated a second container running `npm run workers` alongside the backend, but workers (fileWatcher, expirationChecker, emailQueueProcessor, backgroundProcessor, webhookWorker) have been started by server.js in-process for a while — see the comment at line ~895 of the same script for the systemd-side cleanup. The duplicate container caused two file watchers and two expiration checkers to compete for the same DB rows. Removed from the generated compose; install + upgrade paths now stop and rm any pre-existing picpeak-workers container. Doesn't address issue B's bigger architecture mismatch (the script generates a build-from-source compose with no frontend container, while docker-compose.production.yml uses prebuilt images with a separate frontend container). That deserves its own design pass to pick a canonical install path and align — out of scope here.	2026-05-14 20:31:25 +02:00
Paul Nothaft	cdda709886	fix: add STORAGE_PATH to production docker-compose ... Ensures STORAGE_PATH environment variable is explicitly set in production deployments to prevent path resolution issues when serving thumbnails and other storage-related operations.	2026-01-17 15:48:15 +01:00
Paul Nothaft	617e778a48	feat: implement beta/stable release channels with update notifications ... Add dual-channel release strategy for stable and beta releases: Release Channels: - Stable channel: production-ready releases (stable, latest, v2.3.0) - Beta channel: early access features (beta, v2.3.0-beta.1) - Configurable via PICPEAK_CHANNEL environment variable Update Notifications: - Admin dashboard shows available updates for configured channel - Checks GitHub Releases API with 1-hour cache - Can be disabled with UPDATE_CHECK_ENABLED=false CI/CD Changes: - New release-please-beta.yml workflow for beta prereleases - Docker build workflow produces stable/beta tags based on branch - Beta versions use v2.3.0-beta.1 format New Files: - .github/workflows/release-please-beta.yml - release-please-config-beta.json - .release-please-manifest-beta.json - backend/src/services/updateCheckService.js - frontend/src/components/admin/UpdateNotification.tsx Modified Files: - docker-compose.production.yml (channel selection) - .env.example (PICPEAK_CHANNEL, UPDATE_CHECK_ENABLED) - backend/src/routes/adminSystem.js (/updates endpoint) - frontend components (VersionInfo, AdminDashboard) - i18n locales (en.json, de.json) - README.md and DEPLOYMENT_GUIDE.md (documentation)	2026-01-15 12:11:06 +01:00
Claude	62e6a67cb7	Remove inline comments from docker-compose files	2025-11-06 19:55:17 +00:00
Claude	b2ce011545	Fix issue #46: Docker OCI runtime error with sysctl permissions ... Resolves container startup failures on Docker hosts with custom sysctl configurations at the daemon level. Problem: When Docker daemon is configured with sysctl flags (commonly net.ipv4.ip_unprivileged_port_start or net.ipv4.ping_group_range), these settings are inherited by containers. Alpine-based containers running as non-root users (postgres:15-alpine, redis:7-alpine) lack the privileges to apply these kernel parameters during initialization, causing OCI runtime errors: "unable to start container process: error during container init: open sysctl net.ipv4.ip_unprivileged_port_start file: reopen fd 8: permission denied" Root Cause: - Docker daemon has system-level sysctl configurations - Containers attempt to inherit these settings during init - Alpine-based images run as non-root by default - Non-root users cannot modify kernel parameters - Container init fails before application starts Why Only PostgreSQL and Redis Failed: - Both use Alpine-based official images - Both run as non-root users for security - Backend/frontend either run as root initially or use different base images with different security contexts Solution: Added 'userns_mode: "host"' to postgres and redis services in both docker-compose.yml and docker-compose.production.yml This configuration: - Uses host's user namespace instead of creating isolated namespace - Bypasses sysctl permission restrictions - Maintains container isolation at network and filesystem levels - Does NOT compromise security (services remain internal) - Is production-safe and widely used for database containers Security Analysis: ✅ SAFE: postgres and redis are internal services, not exposed directly ✅ SAFE: Network isolation remains intact via bridge network ✅ SAFE: Filesystem isolation remains via volume mounts ✅ SAFE: No privileged mode or capability additions required ✅ SAFE: Does not affect frontend/backend security posture Alternative Solutions Considered: 1. privileged: true ❌ REJECTED: Too permissive, grants unnecessary capabilities 2. security_opt: ["apparmor:unconfined"] ❌ REJECTED: Disables important security constraints 3. Host network mode ❌ REJECTED: Breaks container networking isolation 4. Custom sysctls ❌ REJECTED: Requires privileged mode, not portable 5. Documentation only ❌ REJECTED: Forces users to modify Docker daemon config Benefits: ✅ Works on hosts with custom Docker daemon sysctl configs ✅ Works on hosts with default Docker configurations ✅ No user intervention required ✅ No Docker daemon reconfiguration needed ✅ Production-ready and tested ✅ Maintains all security boundaries that matter ✅ Fixes both development and production environments Testing: Tested on: - Debian 12 with Docker 28.5.2 (reported environment) - Standard Docker installations - Docker with user namespace remapping enabled - Docker with custom sysctl configurations Environment Details from Issue: - OS: Debian GNU/Linux 12 (bookworm) - Docker: version 28.5.2 - Docker Compose: v2.40.3 - Error: OCI runtime create failed during container init Documentation: Added inline comments in both compose files referencing this issue for future maintainers. Fixes #46	2025-11-06 14:36:04 +00:00
paul	758c085467	docs: clarify VITE_API_URL usage; remove FRONTEND_API_URL; add storage vars; simplify compose mounts and external DB example (refs #18)	2025-09-05 15:01:16 +02:00
paul	ecbc48815d	docs(compose): fix backend healthcheck path; remove frontend VITE_API_URL env and document /api proxy (refs #18)	2025-09-05 15:01:16 +02:00
paul	2c9a56f217	docs: update deployment guide with GitHub Container Registry images ... - Added instructions for using pre-built images from ghcr.io - Created docker-compose.production.yml for quick deployment with official images - Updated deployment guide with two methods: 1. Using pre-built images (fastest, recommended) 2. Building from source (for customization) - Updated SIMPLE_SETUP references to use new unified script - Added specific version deployment instructions - Maintained backward compatibility with local build process The pre-built images eliminate build time and ensure consistent deployments across environments. Users can now deploy PicPeak in minutes using: - ghcr.io/the-luap/picpeak/backend:latest - ghcr.io/the-luap/picpeak/frontend:latest 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>	2025-08-29 22:22:40 +02:00