opensource/mateclaw
2
0
Fork 0
mirror of https://gitee.com/mateos/mateclaw.git synced 2026-09-15 11:58:34 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

build(docker): install poppler + tesseract for PDF extraction parity with Mac

Browse Source
This commit is contained in:
matevip 2026-04-25 11:05:29 +08:00
parent 53f54fe05d
commit 726265780e
1 changed files with 11 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
mateclaw-server/Dockerfile
View File

@ -66,23 +66,26 @@ WORKDIR /app
# #
# PDF extraction toolchain — DocumentExtractTool tries pdftotext first, then # PDF extraction toolchain — DocumentExtractTool tries pdftotext first, then
# Python pdfplumber/pypdf, then falls through to a naive Java parser that # Python pdfplumber/pypdf, then falls through to a naive Java parser that
# reads bytes as ISO_8859_1 (mojibake for CJK). Without poppler-utils and # reads bytes as ISO_8859_1 (mojibake for CJK). Without poppler-utils the
# python3, Docker always hits the naive path and feeds garbled text to the # Docker image always hits the naive path and feeds garbled text to the
# Wiki pipeline. Installing these puts the Docker runtime on the same # Wiki pipeline.
# extraction path as local Mac (brew poppler + pyenv pdfplumber). #
# Tesseract + chi-sim is the final OCR fallback for scanned PDFs. # We install poppler-utils (backend 1) and tesseract (backend 4), which
# together cover the vast majority of PDFs including scanned docs. The
# Python backend is intentionally skipped — pip install against aliyun
# mirrors in CN networks hits transient hash-mismatch failures on cffi /
# cryptography transitive deps, and RFC-051 PR-1c will replace the Python
# hop with JVM-native Tika extraction anyway. Leaving it out keeps the
# image ~200 MB smaller and the build reproducible.
RUN apt-get update \ RUN apt-get update \
&& apt-get install -y --no-install-recommends \ && apt-get install -y --no-install-recommends \
openjdk-21-jre-headless \ openjdk-21-jre-headless \
fonts-noto-cjk \ fonts-noto-cjk \
fonts-noto-color-emoji \ fonts-noto-color-emoji \
poppler-utils \ poppler-utils \
python3 \
python3-pip \
tesseract-ocr \ tesseract-ocr \
tesseract-ocr-chi-sim \ tesseract-ocr-chi-sim \
tzdata \ tzdata \
&& pip3 install --no-cache-dir --break-system-packages pdfplumber pypdf \
&& rm -rf /var/lib/apt/lists/* && rm -rf /var/lib/apt/lists/*
# Tell Playwright Java where Microsoft's image stored the browsers. # Tell Playwright Java where Microsoft's image stored the browsers.