ML-KEM: break out the tests into encap and seeding

[dingxiangfei2009]

cc @davidben

I inspected the these test vector files and I suppose that it is preferable to have tests with similar procedures in their own files.

Fix #191

[sgmenda]

@cpu can we merge this? as @davidben noted in #191 (comment) it would simplify parsing and be in-line with the ML-DSA vectors. also I don't see any objections in #191.

[cpu]

Sorry for letting this one sit.

I talked to Filippo recently and he convinced me that we should go ahead and do the split, but also mentioned that this PR's filenames don't seem right and that we should do a pass to make sure notes & other fields are only present in the files that use them. I think we should address those points before merging.

Separately CI should also enforce the unique type per-test file property & we should describe that in the contrib docs, but neither need to be a blocker here.

[cpu]

I think we should address those points before merging.

I went ahead and recreated this PR here: #199 I'm going to close this iteration out since mine also has the contribution docs update + linting.

Thanks for getting this started dingxiangfei2009!