Pin protobuf version for tensoflow

[xcompass]

Describe your changes

TensorFlow 2.20.0 was built using protobuf version 5.28.3. And currently protobuf 6.31.1 was installed in the image. Then running TF, the following UserWarning messages are printed to users:

UserWarning: Protobuf gencode version 5.28.3 is exactly one major version older than the runtime version 6.31.1 at
tensorflow/core/framework/attr_value.proto. Please update the gencode to avoid compatibility violations in the next runtime release.

Related issue from upstream: tensorflow/tensorflow#98980

Issue ticket if applicable

Checklist (especially for first-time contributors)

• I have performed a self-review of my code
• If it is a core feature, I have added thorough tests
• I will try not to use force-push to make the review process easier for reviewers
• I have updated the documentation for significant changes

[mathbunnyru]

@xcompass I think you didn’t notice, but the tests are failing
You need to add the pin to the exclude list here:

docker-stacks/tests/by_image/docker-stacks-foundation/test_packages.py

Line 50 in 118b7e4

EXCLUDED_PACKAGES = [

[xcompass]

Thanks! I didn't get any notifications about the failing tests. Just pushed a new commit and hopefully it fixes the tests!

[mathbunnyru]

I'm so sorry, @xcompass, it seems our test suite doesn't work well with complicated expressions for versions.
Please, either fix how the version is being parsed, or we can hardcode the latest version of protobuf, which works well.

In either case, please test locally as well (let me know if you need help to set it up).

[xcompass]

No worries @mathbunnyru . Any pointers how to set up test locally? I'm on mac. Thanks

[mathbunnyru]

@xcompass:

1. Install dev dependencies: https://github.com/jupyter/docker-stacks/blob/main/requirements-dev.txt
2. Build images you want to build. I sometimes do make build-all
3. Run tests on an image: make test/scipy-notebook

A few caveats:

1. If the parent image was not built and is not available locally when you're building some image, it will be downloaded from the registry. Depending on the change, it might be fine, but sometimes it might not (for example, when you want to check howthe change in the base-notebook affected all images)
2. If you want to run commands manually, open Makefile. There is nothing fancy there, but it makes life easier
3. To build cuda enabled images, I'm afraid you'll have to build them manually (since Makefile wasn't updated for them and there were discrepancies between architectures)