diff --git a/Dockerfile b/Dockerfile index bc7c1136..decba2f6 100644 --- a/Dockerfile +++ b/Dockerfile @@ -314,7 +314,8 @@ RUN --mount=type=cache,target=/var/cache/apt,sharing=locked,id=apt-$TARGETARCH$T && export CHROME_USER_DATA_DIR="$LIB_DIR/chrome_profile" \ && mkdir -p "$LIB_DIR" \ && apt-get update -qq \ - && apt-get install -qq -y --no-install-recommends build-essential \ + && apt-get install -qq -y --no-install-recommends build-essential tesseract-ocr tesseract-ocr-eng \ + && abxpkg install --no-cache --binproviders=pip --bin-dir="$LIB_DIR/env/bin" gallery-dl \ && if [ "$TARGETARCH" = "arm64" ]; then \ abxpkg install --binproviders=npm --overrides='{"npm":{"install_args":["playwright@next"]}}' playwright; \ abxpkg install --no-cache --install-timeout=600 --binproviders=playwright --bin-dir="$LIB_DIR/env/bin" chromium; \ @@ -384,7 +385,7 @@ RUN chmod +x "$CODE_DIR"/bin/*.sh \ && chown -R "$DEFAULT_PUID:$DEFAULT_PGID" "$LIB_DIR" \ && chmod g+w "$TMP_DIR" "$LIB_DIR" "$LIB_DIR"/bin "$PLAYWRIGHT_BROWSERS_PATH" \ && TIMEOUT=600 gosu "$ARCHIVEBOX_USER" archivebox install 2>&1 | tee -a /VERSION.txt \ - && TIMEOUT=600 gosu "$ARCHIVEBOX_USER" archivebox install archivewebpage liteparse mercury papersdl readability search_backend_sonic opendataloader search_backend_ripgrep 2>&1 | tee -a /VERSION.txt \ + && TIMEOUT=600 gosu "$ARCHIVEBOX_USER" archivebox install archivewebpage gallerydl liteparse mercury papersdl readability search_backend_sonic opendataloader search_backend_ripgrep 2>&1 | tee -a /VERSION.txt \ && gosu "$ARCHIVEBOX_USER" archivebox version 2>&1 | tee -a /VERSION.txt \ && find /venv "$CODE_DIR" "$LIB_DIR" "$DATA_DIR" -type d -name __pycache__ -prune -exec rm -rf {} + \ && find /venv "$CODE_DIR" "$LIB_DIR" "$DATA_DIR" -type f \( -name '*.pyc' -o -name '*.pyo' \) -delete \ diff --git a/Dockerfile.multistage b/Dockerfile.multistage index f543b711..4cc71653 100644 --- a/Dockerfile.multistage +++ b/Dockerfile.multistage @@ -94,7 +94,7 @@ RUN --mount=type=cache,target=/var/cache/apt,sharing=locked,id=apt-$TARGETARCH$T libpango-1.0-0 libx11-6 libx11-xcb1 libxcb1 libxcomposite1 libxdamage1 \ libxext6 libxfixes3 libxkbcommon0 libxrandr2 libxshmfence1 \ fonts-liberation fonts-noto-color-emoji xdg-utils \ - ffmpeg imagemagick tesseract-ocr openjdk-21-jre-headless \ + ffmpeg imagemagick tesseract-ocr tesseract-ocr-eng openjdk-21-jre-headless \ && rm -rf /var/lib/apt/lists/* # Runtime-owned layers copied from the abx-dl image. @@ -197,7 +197,7 @@ RUN echo "[+] Initializing image collection..." \ RUN chmod +x "$CODE_DIR"/bin/*.sh \ && chown -R "$DEFAULT_PUID:$DEFAULT_PGID" "$LIB_DIR" \ && chmod g+w "$TMP_DIR" "$LIB_DIR" "$PLAYWRIGHT_BROWSERS_PATH" \ - && TIMEOUT=600 gosu "$ARCHIVEBOX_USER" archivebox install liteparse mercury papersdl readability search_backend_ripgrep search_backend_sonic 2>&1 | tee -a /VERSION.txt \ + && TIMEOUT=600 gosu "$ARCHIVEBOX_USER" archivebox install gallerydl liteparse mercury papersdl readability search_backend_ripgrep search_backend_sonic 2>&1 | tee -a /VERSION.txt \ && "$LIB_DIR/env/bin/chromium" --version | tee -a /VERSION.txt \ && "$LIB_DIR/pip/packages/papers-dl/venv/bin/papers-dl" --version | tee -a /VERSION.txt \ && /usr/bin/rg --version | head -1 | tee -a /VERSION.txt \