Dockerfile.job

# Job image for the local podman simulation: the deps of BOTH jobs preinstalled so
# the simulated job can skip `pip install` each run (CC_PODMAN_SKIP_PIP=1). Mirrors
# what the real HF Jobs install at runtime into python:3.12 — the estimate job needs
# only duckdb + huggingface_hub (its index SQL is vendored), the fetch job needs
# cdx_toolkit[hf] + uvloop.
FROM python:3.12-slim
RUN apt-get update && apt-get install -y --no-install-recommends git \
    && rm -rf /var/lib/apt/lists/*
RUN pip install --no-cache-dir duckdb huggingface_hub uvloop \
    "cdx_toolkit[hf] @ git+https://github.com/commoncrawl/cdx_toolkit.git@feat/warc-range-sources"