Dockerfile

# Use official Python image
FROM python:3.10

# Set working directory
WORKDIR /

# Copy project files
COPY . .

# Define writable cache locations
# Setup all huggingface and nltk cache dirs to writable /tmp
ENV HF_HOME=/tmp/hf_home
ENV TRANSFORMERS_CACHE=/tmp/hf_home
ENV HUGGINGFACE_HUB_CACHE=/tmp/hf_home
ENV HF_HUB_ENABLE_HF_TRANSFER=0
ENV HF_HUB_DISABLE_SYMLINKS_WARNING=1
ENV HF_HUB_DISABLE_XET=1
ENV NLTK_DATA=/tmp/nltk_data

# Create required directories and ensure proper permissions
RUN mkdir -p /tmp/hf_home/models /tmp/hf_home/hub /tmp/nltk_data /tmp/reports && \
    chmod -R 777 /tmp/hf_home /tmp/nltk_data /tmp/reports


# Optional: expose as env variable
ENV REWRITTEN_OUTPUTS_DIR=/tmp/rewritten_outputs



# Install dependencies
RUN pip install --upgrade pip
RUN pip install -r requirements.txt

# Download NLTK tokenizer
RUN python -m nltk.downloader punkt -d /tmp/nltk_data

# Expose port
EXPOSE 7860

# Start FastAPI server
CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]