Dockerfile

[michabbb]

hi,

i am not a python guy and i was not able to create working docker image.
maybe someone can help here, thanks !!!

here´s my try, but when starting a container, I always get errors that libraries are not found, although they got installed ;(

FROM ubuntu:24.10

ENV DEBIAN_FRONTEND=noninteractive

RUN apt-get update && \
    apt-get install -y \
    python3.10 \
    python3-pip \
    python3-venv \
    git \
    wget \
    libjpeg-dev \
    zlib1g-dev \
    libopencv-dev \
    libgl1 \
    && apt-get clean

WORKDIR /app

RUN git clone https://github.com/opendatalab/PDF-Extract-Kit.git

WORKDIR /app/PDF-Extract-Kit

RUN echo "matplotlib\nPyMuPDF\nultralytics\npaddlepaddle\npaddleocr==2.7.3\npaddlepaddle-gpu" > requirements.txt

RUN python3 -m venv venv && \
    . venv/bin/activate && \
    pip install --upgrade pip
RUN  pip install -r requirements.txt --break-system-packages
RUN  pip install --break-system-packages --extra-index-url https://miropsota.github.io/torch_packages_builder detectron2==0.6+pt2.3.1cu121
RUN  pip install pillow>=10.0.0 --break-system-packages
RUN  pip install opencv-python opencv-python-headless --break-system-packages

COPY . .

# Setze den Befehl zum Starten des Containers
CMD ["bash", "-c", "source venv/bin/activate && python3 pdf_extract.py"]

[wufan-tb]

Is the error info caused by cv2 ? In my experiments, these packages are requried by cv2: libgl1-mesa-glx ffmpeg libsm6 libxext6 git ninja-build libglib2.0-0 libsm6 libxrender-dev libxext6.

Besides, the high version of pillow may cause confict with detection2, try a lower version like 8.4.0

[michabbb]

thanks a lot for your feedback. to be honest, I tried more than 2 hours to get this working, I´m done with that 😖
if someone ever provides a working dockerfile, I am happy to test this project, but for now, it already took too much of my time. again, I am not python guy so figuring out solutions for all these version conflicts, takes me forever. and I am pretty sure someone here has the knowledge and experience to make this in a few minutes 🙈 (and yes, cv2 was one of the errors I had to deal with)