meson-python: The poppler package is split between purelib and platlib

[bzamecnik]

pip install python_poppler==0.4.0 on python 3.10 (Ubuntu 22.04) using meson 1.1.0 fails with:

#30 28.35       + meson install --no-rebuild --destdir /tmp/pip-install-d59qdq37/python-poppler_8100653bafcf4616b4f77887e540edc4/.mesonpy-zrqhtx8x/install
#30 28.35       
#30 28.35       meson-python: error: The poppler package is split between purelib and platlib: 'purelib/poppler/__init__.py' and 'platlib/poppler/cpp/global_.cpython-310-x86_64-linux-gnu.so', a "pure: false" argument may be missing in meson.build
#30 28.35       [end of output]

-> "pure: false" argument may be missing in meson.build

Caused by meson-python-0.13.0 release:

• Raise an error when a package is split between platlib and purelib.

More info: #74

I'd guess that we need to pass pure: false in the meson.build file here (similar to test code of meson, release notes, docs):

python3 = python_mod.find_installation('python3', pure: false)

[frivas-at-navteca]

I am having the same issue, it is really awesome there is a fix already. I am wondering when it will be released so the package installed via pip will work.

Thanks for fixing it so fast!

[bzamecnik]

Yeah, it broke a build and needed to be fixed ASAP.

[eRuaro]

Hi @bzamecnik, any idea how to circumvent this issue when trying to install poppler on docker? Here's my dockerfile, and my requirements.txt

FROM python:3.10-slim-buster

# Update package lists
RUN apt-get update && apt-get install ffmpeg libsm6 libxext6 gcc g++ git build-essential libpoppler-cpp-dev pkg-config poppler-utils tesseract-ocr libtesseract-dev -y

# Make working directories
RUN  mkdir -p  /app
WORKDIR  /app

# Copy the requirements.txt file to the container
COPY requirements.txt .

# Install dependencies
RUN pip install --upgrade pip

RUN pip install torch torchvision torchaudio

RUN pip install -r requirements.txt

RUN pip install 'git+https://github.com/facebookresearch/detectron2.git@v0.4#egg=detectron2'

# Copy the .env file to the container
COPY .env .

# Copy every file in the source folder to the created working directory
COPY  . .

# Expose the port that the application will run on
EXPOSE 8000

# Start the application
CMD ["python", "-m", "uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8000"]

fastapi
uvicorn
langchain
python-poppler
pytesseract
unstructured[local-inference]
psycopg2-binary
pgvector
openai
tiktoken
python-dotenv

[frivas-at-navteca]

Hello @eRuaro

Great question. I hope this answer helps you somehow.

In my specific case I did the following:

FROM public.ecr.aws/lambda/python:3.9
...
RUN pip3 install 'git+https://github.com/facebookresearch/detectron2.git'
...
RUN yum install tesseract tesseract-langpack* poppler poppler-cpp poppler-utils -y
RUN pip3 install --no-cache-dir --upgrade deepdoctection[pt]
RUN pip3 install 'git+https://github.com/mindee/doctr.git'

My suggestion would be trying installing from the source, something like this:

RUN pip3 install 'git+https://github.com/cbrunet/python-poppler.git'

I can see that the main already includes the fix so I would think the new package will be ready in short.

[cbrunet]

Fixed by #75