Issue with Importing fitz (PyMuPDF) in AWS Lambda

[OKihichak]

Hi everyone,

I'm trying to use fitz (PyMuPDF) in an AWS Lambda function, but I keep running into the following error:

ERROR] Runtime.ImportModuleError: Unable to import module 'lambda_function': cannot import name '_extra' from partially initialized module 'pymupdf' (most likely due to a circular import) (/opt/python/pymupdf/__init__.py)

I've tried adding pymupdf to a Lambda layer, but it seems to cause issues due to its dependencies being binary-based. I saw that in the past, people used AWS CodeCommit to resolve this, but now it appears that creating a new CodeCommit repository is no longer possible.

Does anyone know the best way to get fitz running inside AWS Lambda

Any help or guidance would be greatly appreciated! Thanks in advance. 🚀

[gsnoopy]

Do you tried use docker? You need to create an environment that supports binary dependencies.

Create a dockerfile

FROM public.ecr.aws/lambda/python:3.9

RUN yum install -y gcc-c++ make cmake \
    && yum install -y poppler poppler-utils poppler-devel \
    && yum install -y ghostscript

WORKDIR /lambda

RUN pip install --target=/lambda/python pymupdf

CMD ["bash"]

Build

docker build -t pymupdf-layer .

Generate ZIP

docker run --rm -v $(pwd):/out pymupdf-layer bash -c "cd /lambda && zip -r /out/pymupdf-layer.zip python"

Upload zip

aws lambda publish-layer-version \ --layer-name pymupdf-layer \ --description "AWS Lambda layer for PyMuPDF (fitz)" \ --zip-file fileb://pymupdf-layer.zip \ --compatible-runtimes python3.9

• Go to AWS Lambda Console
• Open your Lambda function
• In the Layers section, click Add a layer
• Select Custom layers and choose the pymupdf-layer you just uploaded
• Click Add

[OKihichak]

Thank you so much, I tried it and now it works!

[julian-smith-artifex-com]

This might be a duplicate of #430.