# Seeds a Lance dataset so DuckDB has something to read. The catalog records
# where a table lives and does not carry its data, so the writing half is
# pylance; DuckDB then reads it over S3 with no catalog involved.
FROM python:3.11-slim

WORKDIR /app

RUN pip install --no-cache-dir \
    "lance-namespace==0.8.6" \
    "pylance==10.0.0" \
    "pyarrow==25.0.1"

COPY seed_table.py /app/

CMD ["python3", "/app/seed_table.py", "--help"]