Dockerfile

# Use official slim Python image
FROM python:3.10-slim

# Set environment variables to control cache location
ENV TRANSFORMERS_CACHE=/app/hf_cache \
    HF_HOME=/app/hf_cache \
    XDG_CACHE_HOME=/app/hf_cache \
    TORCH_HOME=/app/hf_cache \
    HF_DATASETS_CACHE=/app/hf_cache \
    SAFE_TENSORS_CACHE=/app/hf_cache

# Set working directory
WORKDIR /app

# Install git for loading models from Hugging Face
RUN apt-get update && apt-get install -y git && rm -rf /var/lib/apt/lists/*

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

# Pre-download the model and tokenizer (build time caching)
RUN python -c "\
from transformers import AutoTokenizer, AutoModelForSequenceClassification; \
model_name = 'tabularisai/multilingual-sentiment-analysis'; \
AutoTokenizer.from_pretrained(model_name); \
AutoModelForSequenceClassification.from_pretrained(model_name); \
"

# Copy application code
COPY . .

# Expose port for Uvicorn
EXPOSE 7860

# Command to run FastAPI with Uvicorn
CMD ["uvicorn", "sentiment_api:app", "--host", "0.0.0.0", "--port", "7860"]