Dockerfile

FROM python:3.9.20-bookworm

# Show print logs; don't write .pyc files.
ENV PYTHONUNBUFFERED=1
ENV PYTHONDONTWRITEBYTECODE=1

# Update the package list
RUN apt update

# Install Gunicorn for production image
RUN pip install gunicorn

# Install project dependencies
COPY requirements.txt /
RUN pip install --no-cache-dir -r /requirements.txt

# transformers 4.20.1 (from requirements.txt) has issues downloading models
# from Hugging Face. This higher version works, but it is incompatible with
# some other dependencies, so we install it here.
RUN pip install --no-cache-dir --upgrade transformers==4.35.0

# Copy the source code
COPY app /app

# Copy model weights
COPY model.tar.gz /app/model.tar.gz

# Set the working directory
WORKDIR /app

# Set application variables for Flask
ENV FLASK_APP=app.py
ENV FLASK_DEBUG=0
ENV MG_PARSER_PORT=32770

# Expose the port on which the application runs.
EXPOSE $MG_PARSER_PORT

# Shell form (instead of recommended JSON form) to allow variable expansion.
CMD flask run --host=0.0.0.0 --port=${MG_PARSER_PORT}