[FEAT]: Interval dtype

[universalmind303]

Description

implement the arrow interval datatype in daft.

Why?

Our current duration dtype is useful for various operations, but it only designed to handle fixed durations. Users will often want to perform operations using relative times such as years, months, and weeks, but this is currently only possible by incorporating additional libraries to handle the relative to fixed conversions for us.

The addition of Interval means that we can natively operate on relative times. This opens up a lot of possibilities for SQL, as we can now use the SQL INTERVAL operator to operate on datetimes.

Still TODO

• implement ops for Interval
  • add
  • sub
• python tests

Note for reviewer:

arrow's interval spec is kinda weird, and is actually 3 different implementations depending on the interval

• years and months are represented as a logical type over i32
• days/millis is a physical type over (i32, i32)
• month/day/nano is a physical type over (i32, i32, i64)

Since most of the kernels are built around the month/day/nano variant, I created the daft Interval type around that one. This introduces a slight amount of overhead when using only year/month (extra i32 & i64) or days/millis (extra i64)

But considering the most common pattern for interval is to do date math on Timestamp columns, this should not be a huge concern.

[codspeed-hq]

CodSpeed Performance Report

Merging #3018 will not alter performance

_{Comparing universalmind303:interval-dtype (51fda30) with main (c2397bf)}

Summary

✅ 17 untouched benchmarks

[universalmind303]

these match statements were very difficult to comprehend. the comments IMO make it much easier to follow the matching logic.

[codecov]

Codecov Report

Attention: Patch coverage is 72.26463% with 109 lines in your changes missing coverage. Please review.

Project coverage is 78.38%. Comparing base (0c39007) to head (51fda30).
Report is 4 commits behind head on main.

Files with missing lines	Patch %	Lines
src/daft-core/src/datatypes/interval.rs	80.00%	46 Missing ⚠️
src/daft-core/src/series/ops/arithmetic.rs	38.46%	16 Missing ⚠️
src/daft-core/src/array/ops/repr.rs	0.00%	10 Missing ⚠️
src/daft-core/src/array/from_iter.rs	50.00%	9 Missing ⚠️
src/daft-core/src/array/serdes.rs	0.00%	9 Missing ⚠️
src/daft-dsl/src/lit.rs	56.25%	7 Missing ⚠️
src/daft-core/src/series/serdes.rs	0.00%	4 Missing ⚠️
src/daft-schema/src/python/datatype.rs	0.00%	3 Missing ⚠️
src/daft-core/src/array/ops/sort.rs	0.00%	2 Missing ⚠️
src/daft-table/src/repr_html.rs	0.00%	2 Missing ⚠️
... and 1 more

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3018      +/-   ##
==========================================
- Coverage   78.39%   78.38%   -0.02%     
==========================================
  Files         603      604       +1     
  Lines       71443    71897     +454     
==========================================
+ Hits        56007    56355     +348     
- Misses      15436    15542     +106     

Files with missing lines	Coverage Δ
daft/expressions/__init__.py	100.00% <100.00%> (ø)
daft/expressions/expressions.py	93.78% <100.00%> (+0.02%)	⬆️
src/daft-core/src/array/growable/arrow_growable.rs	100.00% <ø> (ø)
src/daft-core/src/array/growable/mod.rs	100.00% <ø> (ø)
src/daft-core/src/array/ops/as_arrow.rs	100.00% <ø> (ø)
src/daft-core/src/array/ops/compare_agg.rs	66.50% <ø> (ø)
src/daft-core/src/array/ops/get.rs	78.57% <ø> (ø)
src/daft-core/src/array/ops/take.rs	99.41% <ø> (ø)
src/daft-core/src/array/ops/time.rs	85.63% <100.00%> (+1.26%)	⬆️
src/daft-core/src/datatypes/infer_datatype.rs	86.00% <100.00%> (+0.08%)	⬆️
... and 19 more

... and 4 files with indirect coverage changes