Spaces:

kshitiz14
/

product_attribute

Sleeping

Update Dockerfile

a77d489 verified 4 months ago

1.51 kB

# Product Attribute Extraction API - Docker image
# Compatible with Hugging Face Docker Spaces (port 7860, non-root user)
# and with plain local `docker run` / docker-compose.
FROM python:3.12-slim

# HF Spaces containers run as a non-root user; create one matching their
# convention (uid 1000) so file permissions work whether this builds on
# HF Spaces or locally.
RUN useradd -m -u 1000 user
WORKDIR /app

# System deps kept minimal; scikit-learn wheels are pure-python/precompiled
# so no build-essential is needed on slim.
COPY --chown=user requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy project
COPY --chown=user data/ ./data/
COPY --chown=user src/ ./src/
COPY --chown=user app.py .

USER user

# Build the labeled dataset and train the model at image build time, so the
# container starts up with a ready-to-serve model.joblib baked in (no
# training step needed at runtime/startup).
RUN cd data && python3 build_dataset.py && cd ../src && python3 train.py

# Hugging Face Docker Spaces require the app to listen on port 7860.
EXPOSE 7860

# Basic container healthcheck against the /health endpoint.
HEALTHCHECK --interval=30s --timeout=5s --start-period=10s --retries=3 \
  CMD python3 -c "import urllib.request; urllib.request.urlopen('http://localhost:7860/health')" || exit 1

# app.py at the repo root is the entrypoint (HF Spaces convention); it
# wires up the extraction pipeline living in src/.
CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]


	# Product Attribute Extraction API - Docker image
	# Compatible with Hugging Face Docker Spaces (port 7860, non-root user)
	# and with plain local `docker run` / docker-compose.
	FROM python:3.12-slim

	# HF Spaces containers run as a non-root user; create one matching their
	# convention (uid 1000) so file permissions work whether this builds on
	# HF Spaces or locally.
	RUN useradd -m -u 1000 user
	WORKDIR /app

	# System deps kept minimal; scikit-learn wheels are pure-python/precompiled
	# so no build-essential is needed on slim.
	COPY --chown=user requirements.txt .
	RUN pip install --no-cache-dir -r requirements.txt

	# Copy project
	COPY --chown=user data/ ./data/
	COPY --chown=user src/ ./src/
	COPY --chown=user app.py .

	USER user

	# Build the labeled dataset and train the model at image build time, so the
	# container starts up with a ready-to-serve model.joblib baked in (no
	# training step needed at runtime/startup).
	RUN cd data && python3 build_dataset.py && cd ../src && python3 train.py

	# Hugging Face Docker Spaces require the app to listen on port 7860.
	EXPOSE 7860

	# Basic container healthcheck against the /health endpoint.
	HEALTHCHECK --interval=30s --timeout=5s --start-period=10s --retries=3 \
	CMD python3 -c "import urllib.request; urllib.request.urlopen('http://localhost:7860/health')" \|\| exit 1

	# app.py at the repo root is the entrypoint (HF Spaces convention); it
	# wires up the extraction pipeline living in src/.
	CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]