wgsl: Add sparseMatrixRange to FPTraits API

[zoddicus]

Fixes #3174

---

Requirements for PR author:

• All missing test coverage is tracked with "TODO" or .unimplemented().
• New helpers are /** documented */ and new helper files are found in helper_index.txt.
• Test behaves as expected in a WebGPU implementation. (If not passing, explain above.)

Requirements for reviewer sign-off:

• Tests are properly located in the test tree.
• Test descriptions allow a reader to "read only the test plans and evaluate coverage completeness", and accurately reflect the test code.
• Tests provide complete coverage (including validation control cases). Missing coverage MUST be covered by TODOs.
• Helpers and types promote readability and maintainability.

When landing this PR, be sure to make any necessary issue status updates.

[zoddicus]

Another stacked change, all the new code is in 744e31b

[jiangzhaoming]

Looks good to me, thanks

[ben-clayton]

LGTM, however I'm not sure if the cache should be re-generated or not based on the code changes. It might be worth checking whether you get the same diffs if we all agree on a NodeJS version to use. See #3187

[ben-clayton]

For this and the other changes, a comment describing what cases holds (non-exhaustive fields) would be good as this made me go cross-eyed trying to figure out what it was actually doing

[zoddicus]

Do you think something like

// Cases: [f32|f16]_[non_]const

is sufficient?

This is what jzm has included for the usage of this pattern that they wrote, but I didn't replicate consistantly.

[zoddicus]

LGTM, however I'm not sure if the cache should be re-generated or not based on the code changes. It might be worth checking whether you get the same diffs if we all agree on a NodeJS version to use. See #3187

I am currently using 20.9.0, which I believe is the newest LTS release and the same one that you pin in that PR.

This PR shouldn't be changing the actual expectations significantly (There are a couple of tests that were using the wrong generator that will be changing, but most should be a no-op). It does change the code in floating_point.ts and the various .spec.ts files, so I think they should have a different hash.

[ben-clayton]

SGTM then, thanks!