Improve build performance through improving artifact patterns

[danyeaw]

Performance: Optimize build time after hatchling migration as pointed out by @RRosio

Following our migration to hatchling as the build backend, build times nearly doubled. This PR implements several optimizations that significantly improve build performance:

• Relocate individual files (non-globbed) from artifacts to force-include
• Streamline glob patterns in artifacts

Performance improvements:

• python -m build -w: 60s → 32s (47% reduction)

[RRosio]

Hi @danyeaw, I tried running the Linux JS Tests locally and the first thing I noticed was that the failure in the CI (and replicated locally) was very similar to the error that I saw on main when I forgot to run npm run build. This leads wonder if the client-side JS files available once the tests are running.. I will take a closer look!

[danyeaw]

Hey @RRosio, do you have instructions for getting the dev environment set up to run the js tests? Thanks!

[RRosio]

Hey @RRosio, do you have instructions for getting the dev environment set up to run the js tests? Thanks!

Hi Dan, yes sorry I forgot to mention that, I followed the documentation here: https://github.com/jupyter/nbclassic/blob/b36454a787b4841531029741dbec93aa3be265b4/CONTRIBUTING.rst#javascript-tests

To verify the steps outlined above, I ran the tests on main, with the additional copying of the security_deprecated.js script over to base/js/security.js, they did pass. I thought that maybe the build step for notebook was not getting picked but when I looked into the files in the wheel and even in my source code, I can see the file is there.

[danyeaw]

The only test failure is a flaky test. @RRosio, I split off the change to use build.js to another PR. I think this one is ready 👍