feat(sql): Adds JsonScanBuilder to daft-scan and read_json to daft-sql

[rchowell]

Description

This PR follows the read_csv and read_parquet pattern for adding the table-value function to SQL. This issue solves/addresses #3196 but there a couple TODOs which could be tackled in later PRs.

• support for schema as an argument
• support for mixing named and positional arguments, now only path is positional and all else must be named.
• argument validation, the python API does some validation that's missing on the SQL side.

Issues

• SQL: add read_json function #3196

[codspeed-hq]

CodSpeed Performance Report

Merging #3683 will not alter performance

_{Comparing rchowell/df-253-sql-add-read_json-function (74d4104) with main (feab49a)}

Summary

✅ 27 untouched benchmarks

[codecov]

Codecov Report

Attention: Patch coverage is 58.82353% with 49 lines in your changes missing coverage. Please review.

Project coverage is 77.63%. Comparing base (feab49a) to head (74d4104).
Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
src/daft-scan/src/builder.rs	26.86%	49 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3683      +/-   ##
==========================================
- Coverage   77.82%   77.63%   -0.19%     
==========================================
  Files         728      729       +1     
  Lines       89919    90290     +371     
==========================================
+ Hits        69975    70098     +123     
- Misses      19944    20192     +248     

Files with missing lines	Coverage Δ
src/daft-sql/src/planner.rs	74.49% <100.00%> (ø)
src/daft-sql/src/table_provider/mod.rs	54.54% <100.00%> (ø)
src/daft-sql/src/table_provider/read_csv.rs	90.90% <ø> (ø)
src/daft-sql/src/table_provider/read_json.rs	100.00% <100.00%> (ø)
src/daft-scan/src/builder.rs	37.32% <26.86%> (-3.12%)	⬇️

... and 9 files with indirect coverage changes

[universalmind303]

not necessary in this PR, but I think it'd be nice to also support an array of paths here (we support this in the dataframe API)

something like

select * from read_json(['./file1.json', './file2.json'])

[rchowell]

Here's an issue for all three to support the array of paths.

#3686