| Marc Kupietz | 9baa27a | 2025-11-29 15:32:16 +0100 | [diff] [blame] | 1 | # Slim version without GermaLemma (saves ~180MB) |
| 2 | # Multi-stage Docker build for size optimization |
| 3 | FROM python:3.12-slim-bookworm AS builder |
| 4 | |
| 5 | # Install build dependencies |
| 6 | RUN apt-get update && apt-get install -y \ |
| 7 | gcc \ |
| 8 | g++ \ |
| 9 | && rm -rf /var/lib/apt/lists/* |
| 10 | |
| 11 | # Set environment variables |
| 12 | ENV PIP_CACHE_DIR="/tmp/.cache/pip" \ |
| 13 | PYTHONPATH="PYTHONPATH:." |
| 14 | ENV VIRTUAL_ENV=/app/venv |
| 15 | ENV PATH="$VIRTUAL_ENV/bin:$PATH" |
| 16 | |
| 17 | # Set the working directory and copy requirements |
| 18 | WORKDIR /app |
| 19 | |
| 20 | # Install Python dependencies WITHOUT germalemma |
| 21 | RUN python -m venv venv |
| 22 | RUN venv/bin/pip install --upgrade pip wheel thinc spacy |
| 23 | |
| 24 | # Production stage |
| 25 | FROM python:3.12-slim-bookworm AS production |
| 26 | |
| 27 | # Install minimal runtime dependencies |
| 28 | RUN apt-get update && apt-get install -y \ |
| 29 | wget \ |
| 30 | coreutils \ |
| 31 | && rm -rf /var/lib/apt/lists/* \ |
| 32 | && apt-get clean |
| 33 | |
| 34 | # Add non-root user FIRST (before copying files) |
| 35 | RUN groupadd -r appuser && useradd -r -g appuser appuser |
| 36 | |
| 37 | # Copy virtual environment from builder and set ownership immediately |
| 38 | COPY --from=builder --chown=appuser:appuser /app/venv /app/venv |
| 39 | |
| 40 | # Copy application code with correct ownership |
| 41 | COPY --chown=appuser:appuser lib /app/lib |
| 42 | COPY --chown=appuser:appuser systems /app/systems |
| 43 | COPY --chown=appuser:appuser my_utils /app/my_utils |
| 44 | COPY --chown=appuser:appuser download_with_progress.py /app/download_with_progress.py |
| 45 | COPY --chown=appuser:appuser list_spacy_models.py /app/list_spacy_models.py |
| 46 | COPY --chown=appuser:appuser docker-entrypoint.sh /docker-entrypoint.sh |
| 47 | |
| 48 | # Set environment variables |
| 49 | ENV VIRTUAL_ENV=/app/venv |
| 50 | ENV PATH="$VIRTUAL_ENV/bin:$PATH" |
| 51 | ENV PYTHONPATH="PYTHONPATH:." |
| 52 | |
| 53 | # spaCy processing configuration - GermaLemma disabled by default |
| 54 | ENV SPACY_USE_DEPENDENCIES="True" |
| 55 | ENV SPACY_USE_GERMALEMMA="False" |
| 56 | ENV SPACY_PARSE_TIMEOUT="30" |
| 57 | ENV SPACY_MAX_SENTENCE_LENGTH="500" |
| 58 | ENV SPACY_N_PROCESS="10" |
| 59 | ENV SPACY_BATCH_SIZE="2000" |
| 60 | ENV SPACY_CHUNK_SIZE="20000" |
| 61 | |
| 62 | WORKDIR /app |
| 63 | |
| 64 | # Create directories with correct ownership |
| 65 | RUN mkdir -p "/app/logs" "/app/tmp" "/local/models" && \ |
| 66 | chown -R appuser:appuser "/app/logs" "/app/tmp" "/local/models" && \ |
| 67 | chmod +x /docker-entrypoint.sh && \ |
| 68 | chmod +x /app/download_with_progress.py && \ |
| 69 | chmod +x /app/list_spacy_models.py |
| 70 | |
| 71 | # Set temp directories to use app directory instead of system /tmp |
| 72 | ENV TMPDIR="/app/tmp" |
| 73 | ENV TEMP="/app/tmp" |
| 74 | ENV TMP="/app/tmp" |
| 75 | |
| 76 | # Switch to non-root user |
| 77 | USER appuser |
| 78 | |
| 79 | # Define the entry point |
| 80 | ENTRYPOINT ["/docker-entrypoint.sh"] |