[featuring] - add hanning window function

[Vlor999]

Description

This PR implements the Hanning window function (mlx.core.hanning), bringing MLX closer to feature parity with NumPy's signal processing capabilities (numpy.hanning).

Window functions are essential for audio processing and various signal analysis tasks.

Implementation Details

• C++: Implemented hanning in mlx/ops.cpp using existing primitives (arange, cos, multiply, subtract). Used M_PI from <cmath> for precision.
• Python Bindings: Exposed the function via pybind11 in python/src/ops.cpp under mlx.core.
• Documentation: Added docstrings describing the parameters and return values, matching the NumPy behavior.

Test Plan

I have verified the implementation locally against NumPy to ensure numerical accuracy.

Verification script:

import mlx.core as mx
import numpy as np

M = 50
mx_hanning = mx.hanning(M)
np_hanning = np.hanning(M)

# Check for numerical closeness
assert np.allclose(mx_hanning, np_hanning, atol=1e-6), "Mismatch with NumPy implementation"
print("✅ Verification passed against numpy.hanning")

Unit Tests:

Added a test case in python/tests/test_ops.py covering standard usage, edge cases (M=1), and empty input (M=0).

Checklist

Put an x in the boxes that apply.

• I have read the CONTRIBUTING document
• I have run pre-commit run --all-files to format my code / installed pre-commit prior to committing changes
• I have added tests that prove my fix is effective or that my feature works
• I have updated the necessary documentation (if needed)

[angeloskath]

Thanks! I left some comments after these we can merge it.

[Vlor999]

Thanks for the feedback!

I've updated the implementation as discussed:

• Optimization: Switched to the $\sin^2$ identity (square(sin(...))). This reduces the compute graph size and memory usage by avoiding the intermediate arrays needed for the cosine formula.
• Verification: Verified that tests still pass and numerical results remain consistent with NumPy.

Ready for another look! 🚀

[angeloskath]

Looks good thanks!