Allow broadcasting in specialized numba dispatch of AdvancedIncSubtensor

[ricardoV94]

---

📚 Documentation preview 📚: https://pytensor--1272.org.readthedocs.build/en/1272/

[codecov]

Codecov Report

Attention: Patch coverage is 95.34884% with 2 lines in your changes missing coverage. Please review.

Project coverage is 82.00%. Comparing base (89d5366) to head (ec81f12).
Report is 9 commits behind head on main.

Files with missing lines	Patch %	Lines
pytensor/link/numba/dispatch/subtensor.py	95.34%	1 Missing and 1 partial ⚠️

❌ Your patch status has failed because the patch coverage (95.34%) is below the target coverage (100.00%). You can increase the patch coverage or adjust the target coverage.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1272   +/-   ##
=======================================
  Coverage   81.99%   82.00%           
=======================================
  Files         188      188           
  Lines       48608    48617    +9     
  Branches     8688     8690    +2     
=======================================
+ Hits        39857    39866    +9     
  Misses       6586     6586           
  Partials     2165     2165           

Files with missing lines	Coverage Δ
pytensor/link/numba/dispatch/subtensor.py	95.58% <95.34%> (+0.23%)	⬆️

🚀 New features to boost your workflow:

• ❄ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[jessegrabowski]

Can we concoct a new test case that still requires object on both inc and set? Or is the coverage now that good?

[ricardoV94]

There are 3 such tests already:

pytensor/tests/link/numba/test_subtensor.py

Lines 398 to 417 in ec81f12

	(
	np.arange(3 * 4 * 5).reshape((3, 4, 5)),
	rng.poisson(size=(2, 4)),
	([1, 2], slice(None), [3, 4]), # Non-consecutive vector indices
	False,
	True,
	True,
	),
	(
	np.arange(3 * 4 * 5).reshape((3, 4, 5)),
	rng.poisson(size=(2, 2)),
	(
	slice(1, None),
	[1, 2],
	[3, 4],
	), # Mixed double vector index and basic index
	False,
	True,
	True,
	),

pytensor/tests/link/numba/test_subtensor.py

Lines 435 to 441 in ec81f12

	np.arange(3 * 5).reshape((3, 5)),
	rng.poisson(size=(1, 2, 2)), # Same as before, but Y broadcasts
	(slice(1, 3), [[1, 2], [2, 3]]),
	False,
	True,
	True,
	),

Either when array indexes are non-consecutive, or mixed with basic indices