feat(rust): implement decode long stream

[Weixing-Zhang]

• Migrate the required functions to new architecture
• Add le_bytes_to_u64s
• Add decode_physical_u64
• Add decode_logical_u64
• Add decode_long_stream
• Add tests

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 53.36%. Comparing base (6287bb5) to head (5bf7d39).

❗ There is a different number of reports uploaded between BASE (6287bb5) and HEAD (5bf7d39). Click for more details.

HEAD has 1 upload less than BASE

Flag	BASE (6287bb5)	HEAD (5bf7d39)
	2	1

Additional details and impacted files

@@             Coverage Diff             @@
##             main     #597       +/-   ##
===========================================
- Coverage   73.61%   53.36%   -20.25%     
===========================================
  Files          66       12       -54     
  Lines        3286      920     -2366     
  Branches     1461      920      -541     
===========================================
- Hits         2419      491     -1928     
+ Misses        767      358      -409     
+ Partials      100       71       -29     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[nyurik]

@Weixing-Zhang hi, please make sure to look at the Slack message i sent to the #maplibre-tile-format

[nyurik]

one big issue with these implementations is that unlike the first iteration (in the /rust/mlt dir), here we should focus on what is actually being used - i.e. any code we add should parse more MLT files than before

[Weixing-Zhang]

one big issue with these implementations is that unlike the first iteration (in the /rust/mlt dir), here we should focus on what is actually being used - i.e. any code we add should parse more MLT files than before

@nyurik Agreed. I will probably integrate the int decoder into the ID decoding in this PR as well.

[nyurik]

@Weixing-Zhang please take a look at the #687 -- i think it will be easier to create PRs against that branch, so that we can see what changes with each PR. Think of it as a work-in-progress branch that we will later squash once we get all decoding done properly

[Weixing-Zhang]

@nyurik I can pick the work in the rust/mlt-nom/src/v01/stream.rs based on #687.

Perhaps we can close this PR then.

[nyurik]

@Weixing-Zhang ok, i just did some vibe coding with #687 - seems now all decoded snapshots are generated, but i suspect they are far from being in a "good state". I am beginning to think we should just implement the MLT -> MVT encoding, and do the comparison with the original data. This way we can let AI deal with the bugs, and just make it ensure generated MVTs are exactly as the original data... In the mean time - please do a deep dive into the decoders - you have far better understanding of how they are being done than I

[Weixing-Zhang]

The implementation is being done in this PR #692 now.