feat: Add {Expr,Series}.sin

[liamholmes31]

Description

Added sin expr/series methods.

Have tried to follow existing PRs adding e.g. log methods - was a bit hazy on which classes in _compliant I needed to add these to, so apologies if anything is missing!

What type of PR is this? (check all applicable)

• 💾 Refactor
• ✨ Feature
• 🐛 Bug Fix
• 🔧 Optimization
• 📝 Documentation
• ✅ Test
• 🐳 Other

Related issues

• Closes [Enh]: add Expr.sin and Series.sin #3355

Checklist

• Code follows style guide (ruff)
• Tests added
• Documented the changes

[dangotbanned]

@liamholmes31 I see that you removed this, but it was on the right track!

Do you mind if I add a commit with where this should be defined?
The file isn't in the changes so I couldn't use a suggestion

[liamholmes31]

sure, thanks @dangotbanned !

[dangotbanned]

Please excuse the typo 🤦‍♂️
fix: Add ComplianColumn.sin

We should probably document this better!

All the methods that Expr and Series share and use Self can be added to CompliantColumn

narwhals/narwhals/_compliant/column.py

Lines 31 to 32 in 8d93fb1

	class CompliantColumn(Protocol):
	"""Common parts of `Expr`, `Series`."""

So for example, CompliantColumn.log is here, since it works elementwise:

narwhals/narwhals/_compliant/column.py

Line 111 in 8d93fb1

def log(self, base: float) -> Self: ...

But for something like first, we need two separate definitions:

Here we aggregate to a literal value

narwhals/narwhals/_compliant/series.py

Line 137 in 8d93fb1

def first(self) -> PythonLiteral: ...

But here the aggregation is a literal expression

narwhals/narwhals/_compliant/expr.py

Line 141 in 8d93fb1

def first(self) -> Self: ...

[liamholmes31]

ahh OK this makes sense, thanks!! I was thrown by not spotting where log was defined in compliant, should have looked a bit harder!

[liamholmes31]

might need some advice on the failing type check CI stage (see below), looks like the type hint for 'pc_func' may need to be relaxed?

  /home/runner/work/narwhals/narwhals/narwhals/_pandas_like/series.py:1098:70 - error: Argument of type "Overload[(x: pyarrow.compute.FloatScalar, /, *, memory_pool: MemoryPool | None = None) -> (pyarrow.__lib_pxi.scalar.FloatScalar | DoubleScalar), (x: FloatArray, /, *, memory_pool: MemoryPool | None = None) -> (NumericArray[pyarrow.__lib_pxi.scalar.FloatScalar] | NumericArray[DoubleScalar]), (x: Expression, /, *, memory_pool: MemoryPool | None = None) -> Expression]" cannot be assigned to parameter "pc_func" of type "(ChunkedArrayAny) -> ChunkedArrayAny" in function "_apply_pyarrow_compute_func"
    No overloaded function matches type "(ChunkedArrayAny) -> ChunkedArrayAny" (reportArgumentType)

[liamholmes31]

some of the failing downstream stages don't look obviously related to my changes here, are these known failing steps?

[FBruzzesi]

Hey @liamholmes31 thanks for you PR 🚀 Looking forward to finalize it

some of the failing downstream stages don't look obviously related to my changes here, are these known failing steps?

They surely aren't related with the changes, we will address them separately or report the issues downstream 😉

[MarcoGorelli]

might need some advice on the failing type check CI stage (see below), looks like the type hint for 'pc_func' may need to be relaxed?

  /home/runner/work/narwhals/narwhals/narwhals/_pandas_like/series.py:1098:70 - error: Argument of type "Overload[(x: pyarrow.compute.FloatScalar, /, *, memory_pool: MemoryPool | None = None) -> (pyarrow.__lib_pxi.scalar.FloatScalar | DoubleScalar), (x: FloatArray, /, *, memory_pool: MemoryPool | None = None) -> (NumericArray[pyarrow.__lib_pxi.scalar.FloatScalar] | NumericArray[DoubleScalar]), (x: Expression, /, *, memory_pool: MemoryPool | None = None) -> Expression]" cannot be assigned to parameter "pc_func" of type "(ChunkedArrayAny) -> ChunkedArrayAny" in function "_apply_pyarrow_compute_func"
    No overloaded function matches type "(ChunkedArrayAny) -> ChunkedArrayAny" (reportArgumentType)

looks like an issue with pyarrow-stubs, i'd suggest just type-ignoring (and ideally reporting upstream)