fix(ml): install ffmpeg in ml-worker image for video frame sampling
extract_video_frames() calls ffprobe/ffmpeg, which were only installed in the web Dockerfile. Videos route through the ml-worker for WD14 + SigLIP inference, so the tool needs to exist there too. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
+3
-1
@@ -6,11 +6,13 @@ ENV PYTHONDONTWRITEBYTECODE=1 \
|
||||
PIP_NO_CACHE_DIR=1 \
|
||||
ML_MODEL_DIR=/models
|
||||
|
||||
# System deps: Pillow / image handling + psycopg2 build deps (binary wheel usually fine but libpq is needed at runtime on slim)
|
||||
# System deps: Pillow / image handling + psycopg2 runtime libpq + ffmpeg for
|
||||
# video frame sampling (WD14+SigLIP inference on videos calls ffprobe/ffmpeg).
|
||||
RUN apt-get update && apt-get install -y --no-install-recommends \
|
||||
libgl1 \
|
||||
libglib2.0-0 \
|
||||
libpq5 \
|
||||
ffmpeg \
|
||||
&& rm -rf /var/lib/apt/lists/*
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
Reference in New Issue
Block a user