[FEAT] Support Hudi reader

[xushiyan]

• Add basic Hudi APIs
• Add basic Hudi COW reader support

for #2070

[codecov]

Codecov Report

Attention: Patch coverage is 88.86311% with 48 lines in your changes are missing coverage. Please review.

Project coverage is 84.96%. Comparing base (3c43394) to head (9d99688).
Report is 7 commits behind head on main.

❗ Current head 9d99688 differs from pull request most recent head 34ed977. Consider uploading reports for the commit 34ed977 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2011      +/-   ##
==========================================
+ Coverage   84.72%   84.96%   +0.24%     
==========================================
  Files          62       68       +6     
  Lines        6840     7271     +431     
==========================================
+ Hits         5795     6178     +383     
- Misses       1045     1093      +48     

Files	Coverage Δ
daft/__init__.py	24.24% <ø> (ø)
daft/hudi/pyhudi/timeline.py	100.00% <100.00%> (ø)
daft/io/__init__.py	95.65% <100.00%> (+0.19%)	⬆️
daft/io/_hudi.py	100.00% <100.00%> (ø)
daft/hudi/pyhudi/table.py	99.15% <99.15%> (ø)
daft/hudi/pyhudi/utils.py	98.30% <98.30%> (ø)
daft/hudi/pyhudi/filegroup.py	95.00% <95.00%> (ø)
daft/hudi/hudi_scan.py	66.14% <66.14%> (ø)

[jaychia]

Seems pretty good to me overall!

We're happy to keep code in Daft for now in the daft/hudi/pyhudi folder, but we should probably chat about how you envision the split and API between our libraries.

Couple of important things we should verify:

• Does this work with S3? We should try reading a Hudi table from AWS S3 and make sure that works. We do have some examples of integration tests using minio too that you can look at if you want a locally runnable example (see: tests/integration/iceberg)
• Does this work with the full range of Hudi's typesystem?
• We should also throw an error early if there features we don't support and detect that early (e.g. merge-on-read)

[jaychia]

LGTM!!! 🚀 🚀 🚀 🚀