kjbyg

Dockerfile

@@ -1,6 +1,6 @@
 FROM python:3.10-slim
 
-# 1) Set up cache/env paths
+# 1) Set up env so caches live under the non-root user’s home
 ENV HOME="/home/user" \
     HF_HOME="$HOME/.cache/huggingface" \
     HF_HUB_CACHE="$HF_HOME/hub" \
@@ -8,25 +8,24 @@ ENV HOME="/home/user" \
     DOCLING_ARTIFACTS_PATH="$HOME/.cache/docling/models" \
     OMP_NUM_THREADS=2
 
-# 2) Create non-root user
+# 2) Create non-root user early
 RUN useradd -m -u 1000 user
 
-# 3) Install system dependencies
+# 3) Install system deps
 RUN apt-get update && apt-get install -y \
       build-essential git libpoppler-cpp-dev poppler-utils \
       libmagic-dev python3-dev \
     && rm -rf /var/lib/apt/lists/*
 
-# 4) Copy and install Python dependencies
+# 4) Copy & install Python deps
 WORKDIR $HOME/app
-# Note: this COPY refers to ./requirements.txt in YOUR repo, not /home/user/app
 COPY --chown=user:user requirements.txt ./
 RUN pip install --no-cache-dir \
       torch torchvision --extra-index-url https://download.pytorch.org/whl/cpu \
       huggingface_hub accelerate \
     && pip install --no-cache-dir -r requirements.txt
 
-# 5) Prepare cache directories and set ownership
+# 5) Prepare cache dirs under root and chown to user
 RUN mkdir -p \
       $DOCLING_ARTIFACTS_PATH \
       $HF_HUB_CACHE \
@@ -36,15 +35,15 @@ RUN mkdir -p \
       $HOME/.cache/huggingface \
       $HOME/app/model_cache
 
-# 6) Switch to non-root to prefetch Docling models
+# 6) Switch to non-root: prefetch Docling models into $HOME/.cache/docling/models
 USER user
 RUN pip install --no-cache-dir docling \
  && docling-tools models download
 
-# 7) Prefetch the NuExtract model
+# 7) Prefetch NuExtract model into app/model_cache using user-owned HF cache
 RUN python - <<EOF
-from huggingface_hub import snapshot_download
 import os
+from huggingface_hub import snapshot_download
 snapshot_download(
     repo_id="numind/NuExtract-1.5-tiny",
     local_dir=os.path.join(os.getenv("HOME"), "app", "model_cache"),
@@ -53,6 +52,5 @@ snapshot_download(
 EOF
 
 # 8) Copy application code and start
-USER user
 COPY --chown=user:user . .
 CMD ["uvicorn","app:app","--host","0.0.0.0","--port","7860"]