Dockerfile

# ─── Render Free Tier Optimised Dockerfile ───────────────────────────────
# Free tier: 512 MB RAM, shared CPU.
# Key rules:
#   1. CPU-only torch — saves ~800 MB image size vs default CUDA build
#   2. Pre-download ML models at build time — avoids cold-start timeout
#   3. web3 NOT installed — payments run in local ledger mode by default

FROM python:3.11-slim

# Build tools needed by some C-extension packages
RUN apt-get update && apt-get install -y --no-install-recommends \
    gcc g++ curl \
    && rm -rf /var/lib/apt/lists/*

WORKDIR /app

# ── 1. Install CPU-only torch first (must be before requirements.txt) ─────
# Pulling the CPU wheel avoids downloading 2+ GB of CUDA binaries
RUN pip install --no-cache-dir \
    torch==2.1.0 \
    --index-url https://download.pytorch.org/whl/cpu

# ── 2. Install remaining dependencies ─────────────────────────────────────
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# ── 3. Pre-cache ML models into the image layer ───────────────────────────
# Runs once at build time. Cold starts then take ~5s not ~3 min.
RUN python -c "\
from sentence_transformers import SentenceTransformer, CrossEncoder; \
SentenceTransformer('all-MiniLM-L6-v2'); \
CrossEncoder('cross-encoder/ms-marco-MiniLM-L-6-v2'); \
print('Models pre-cached OK') \
"

# ── 4. Copy application ────────────────────────────────────────────────────
COPY . .

RUN mkdir -p data identity

# ── 5. Runtime ────────────────────────────────────────────────────────────
ENV PYTHONPATH=/app
ENV PYTHONUNBUFFERED=1

EXPOSE 8000

HEALTHCHECK --interval=30s --timeout=15s --start-period=90s --retries=3 \
  CMD curl -f http://localhost:${PORT:-8000}/health || exit 1

# Render injects $PORT automatically
CMD ["sh", "-c", "uvicorn main:app --host 0.0.0.0 --port ${PORT:-8000} --workers 1"]