Files
SpecterOps-Nemesis/projects/document_conversion/Dockerfile
T
harmj0y b28d043795 Start of Tika OCR multi-language support
-Start of Tika OCR multi-language support
2025-10-02 16:39:15 +09:00

122 lines
3.7 KiB
Docker

# FROM nemesis-python-base-dev AS base
ARG PYTHON_BASE_DEV_IMAGE=nemesis-python-base-dev
ARG PYTHON_BASE_PROD_IMAGE=nemesis-python-base-prod
FROM ${PYTHON_BASE_DEV_IMAGE} AS base
ARG TIKA_OCR_LANGUAGES="eng"
ARG TIKA_VERSION="3.1.0"
# Install dependencies including multiple Tesseract language packs
RUN apt-get update && \
apt-get install -y --no-install-recommends \
openjdk-17-jre-headless \
wget \
tesseract-ocr \
libpq-dev \
postgresql-client \
binutils \
# Install fonts for better CJK support
fonts-noto-cjk \
fonts-noto-color-emoji && \
# Install Tesseract language packs dynamically
for lang in ${TIKA_OCR_LANGUAGES}; do \
apt-get install -y --no-install-recommends tesseract-ocr-${lang} || echo "Warning: Failed to install tesseract-ocr-${lang}"; \
done && \
# Download Tika server
wget https://archive.apache.org/dist/tika/${TIKA_VERSION}/tika-server-standard-${TIKA_VERSION}.jar -O /tika-server-standard.jar && \
apt-get clean && \
rm -rf /var/lib/apt/lists/*
# Set JAVA_HOME for JRE
ENV JAVA_HOME=/usr/lib/jvm/java-17-openjdk-amd64/jre
# Copy only what's needed for dependency installation first (leveraging Docker cache)
COPY ./projects/document_conversion/poetry.lock ./projects/document_conversion/pyproject.toml /src/projects/document_conversion/
WORKDIR /src/projects/document_conversion
# Copy remaining files
COPY ./libs /src/libs
COPY ./projects/document_conversion /src/projects/document_conversion/
########################
# Development
########################
FROM base AS dev
# Set environment variables together in one layer
ENV PYTHONUNBUFFERED=1 \
PYTHONDONTWRITEBYTECODE=1 \
LOG_LEVEL=DEBUG \
UVICORN_HOST="0.0.0.0" \
UVICORN_PORT=8000 \
UVICORN_RELOAD_DIR="/src/"
RUN poetry install
ENTRYPOINT ["/bin/sh", "-c", " \
poetry run uvicorn document_conversion.main:app \
--host ${UVICORN_HOST} \
--port ${UVICORN_PORT} \
--reload \
--reload-dir ${UVICORN_RELOAD_DIR} \
"]
########################
# Runtime - Bundle the app
########################
FROM base AS bundle
RUN poetry bundle venv --python=/usr/bin/python3 --only=main /venv
########################
# Make the final image
########################
# FROM nemesis-python-base-prod AS prod
FROM ${PYTHON_BASE_PROD_IMAGE} AS prod
ARG TIKA_OCR_LANGUAGES="eng"
# Install dependencies in a single RUN command to reduce layers
RUN apt-get update && \
apt-get install -y --no-install-recommends \
openjdk-17-jre-headless \
wget \
tesseract-ocr \
libpq-dev \
postgresql-client \
binutils \
# Install fonts for better CJK support
fonts-noto-cjk \
fonts-noto-color-emoji && \
# Install Tesseract language packs dynamically
for lang in ${TIKA_OCR_LANGUAGES}; do \
apt-get install -y --no-install-recommends tesseract-ocr-${lang} || echo "Warning: Failed to install tesseract-ocr-${lang}"; \
done && \
apt-get clean && \
rm -rf /var/lib/apt/lists/*
ENV JAVA_HOME=/usr/lib/jvm/java-17-openjdk-amd64/jre \
LOG_LEVEL=INFO \
UVICORN_HOST=0.0.0.0 \
UVICORN_PORT=8000 \
UVICORN_WORKERS=1 \
UVICORN_PROXY_HEADERS=1 \
UVICORN_ACCESS_LOG=false
# Copy only the virtual environment from the bundle stage
COPY --from=bundle /venv /venv
# Copy the Tika JAR from base so we don't download again
COPY --from=base /tika-server-standard.jar /tika-server-standard.jar
# TODO: Re-enable when we're ready for release
# USER nemesis
ENTRYPOINT ["/bin/sh", "-c", "\
/venv/bin/uvicorn \"document_conversion.main:app\" \
--host ${UVICORN_HOST} \
--port ${UVICORN_PORT} \
--workers ${UVICORN_WORKERS} \
--proxy-headers \
--no-access-log \
"]