Fix invalid ONNX graph from None start/end in aten_slice_scatter

[fruet (gabrielfruet)]

aten::slice_scatter produced an invalid ONNX graph whenever start or end came in as None. The lowering unsqueezed both bounds unconditionally, so a None turned into an empty-string input to Unsqueeze. A missing start now defaults to 0 and a missing end to INT64_MAX, a full slice, matching what aten_slice already does.

Changes:

• torch_lib/ops/core.py: build the Slice bounds conditionally, unsqueezing start or end only when they are provided.
• tests/function_libs/torch_lib/extra_opinfo.py: regression sample-input cases for None start, None end, both, and a step-2 variant.

Testing: the existing slice_scatter OpInfo suite now exercises the None bounds.

Fixes #2372

[fruet (gabrielfruet)]

Justin Chu (@justinchuby) Do you mind taking a look?

[Justin Chu (justinchuby)]

Thanks!

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 72.63%. Comparing base (9ef0aac) to head (36fa482).
⚠️ Report is 12 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #3006   +/-   ##
=======================================
  Coverage   72.63%   72.63%           
=======================================
  Files         265      265           
  Lines       32204    32210    +6     
  Branches     3041     3042    +1     
=======================================
+ Hits        23391    23397    +6     
  Misses       7779     7779           
  Partials     1034     1034           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.