Repository navigation
feat(faces): find and describe faces with anytopdf - #227
Merged
Merged
Conversation
Settings › Video › Face engine › anytopdf: its faces plugin finds the faces (built-in YuNet) and its face-id plugin describes them with the SFace file Lens fetches, so a server needs no OpenCV, and faces match ones OpenCV found (0.88-0.99 on the same face). The faces step sends up to 50 frames or pages per run. anytopdf's face index lives only for the run: Lens reads each face's description from it, and who someone is stays in Lens. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01TbMhC5vN23npJjEVxfEApC
adeelahmad
force-pushed
the
claude/project-thread-l382kt
branch
from
October 7, 2026 11:44
8af3c18 to
31086a5
Compare
adeelahmad
marked this pull request as ready for review
October 7, 2026 11:45
…-l382kt # Conflicts: # CHANGELOG.md # docs/configuration.md # fastapi_backend/app/domain/components.py # fastapi_backend/tests/fake_anytopdf.py
An unread upload made the socket reset, so the client sometimes saw "can't be reached" instead of the long answer (red on CI). Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01TbMhC5vN23npJjEVxfEApC
…-l382kt # Conflicts: # CHANGELOG.md # docs/components.md # fastapi_backend/app/domain/components.py # fastapi_backend/app/domain/settings.py
…-l382kt # Conflicts: # fastapi_backend/app/domain/anytopdf.py # fastapi_backend/tests/api/test_anytopdf.py
…-l382kt # Conflicts: # CHANGELOG.md # fastapi_backend/app/domain/anytopdf.py
The merge left the long-answer mode reading the upload twice, so it waited on the client and broke the pipe. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01TbMhC5vN23npJjEVxfEApC
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Requested by Adeel · project thread
Builds on #221 (base adapter, merged).
Description
Before: faces needed OpenCV (YuNet + SFace) or InsightFace.
After: Settings › Video › Face engine › anytopdf. anytopdf's faces plugin finds the faces (built-in YuNet) and its face-id plugin describes them with the same SFace file Lens fetches, so a server needs no OpenCV or Python packages. Its descriptions match OpenCV's for the same face (0.88 and 0.99 cosine on the OpenCV sample photos, 0.15 between different people), so switching engines keeps the people a namespace already knows.
How:
anytopdf.faces()runs one sandboxed, offline convert over up to 50 frames or pages with--recognize-facesand a throwaway face index; Lens reads each face's box and score from the graph and its embedding from that run'sfaces.sqlitesightings at the same box, since the graph never carries one.analyze()andfaces()share one runner. Identity stays in Lens (faces.py). The components fetch anytopdf and only the SFace file for this engine.Motivation and Context
anytopdf integration plan, phase 2: Lens reading anytopdf 0.4.0's faces.
Types of changes
!in the title)Checklist:
make hookschecks it on commit).docs/accordingly.## UnreleasedinCHANGELOG.md.Checked against the real anytopdf 0.4.0 release on Linux x86-64 (lena.jpg, messi5.jpg); backend suite 693 passed locally.
🤖 Generated with Claude Code
https://claude.ai/code/session_01TbMhC5vN23npJjEVxfEApC