Add shap to ML training/testing libraries

[tvami]

I am adding a new package to the container, here are the details.

Resolves #160

What new packages does this PR add to the development image?

• shap

Check List

• I successfully built the container using docker

• I was able to successfully use the new packages.

docker build . -t ldmx/local:temp-tag
docker run -it --rm -v $PWD:$PWD -w $PWD ldmx/local:temp-tag
python3 -c "import shap"
root@1ae77df9afa4:/tmp/dev-build-context# 

So it works

[JYoo001]

Great, thank you!

[tomeichlersmith]

While I'm fine with including this, I do want to point out that the python ecosystem has much better environment handling than building frozen packages into container images.

I would suggest using uproot/awkward instead of PyROOT. The annoyance of switching to a different package will be outweighted by the time saved (uproot/awkward is faster) and the ease of use. Without using PyROOT, you can just

uv init
uv add scikit-hep

and you are good to go with the latest packages which you can mirror across machines by carrying around the uv.lock file that uv creates. https://docs.astral.sh/uv/

[tvami]

Yes it's long overdue to move the UCSB scripts to uproot. But ok for now let's just merge this. Do I need to cut a release for this to work @tomeichlersmith ?

[tomeichlersmith]

Now that's its merged, it will be built into ldmx/dev:main. A new release would build a new tag that would update ldmx/dev:latest.

This will not propagate into ldmx/pro until a new release of ldmx-sw is made after its Dockerfile is updated to point to the updated ldmx/dev image.

[tvami]

OK so I'd like it to be in the ldmx/dev:latest, I'll make a release then