Improve Reshape and Slice folding

[justinchuby]

• Fold the input to reshape if from the output shape we know it can be folded.
• Eliminate no-op slice if the input shape is the same as output shape.

[Copilot]

Pull request overview

This pull request improves constant folding for Reshape operations by materializing shape inputs as constants when the output shape is known from shape inference. It also adds support for INT32 dtype shape inputs, which are common in TFLite models.

Changes:

• Added INT32 fallback support in get_shape_value() for shape inputs (common in TFLite models)
• Implemented _try_materialize_reshape_shape() to create constant shape inputs from inferred output shapes
• Updated reshape() function to leverage shape materialization for optimization

[codecov]

Codecov Report

❌ Patch coverage is 92.48120% with 10 lines in your changes missing coverage. Please review.
✅ Project coverage is 70.61%. Comparing base (2b2618e) to head (e790612).
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
...er/rules/common/_materialize_reshape_shape_test.py	89.61%	1 Missing and 7 partials ⚠️
...nxscript/rewriter/rules/common/_collapse_slices.py	60.00%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2807      +/-   ##
==========================================
+ Coverage   70.50%   70.61%   +0.11%     
==========================================
  Files         228      230       +2     
  Lines       27114    27246     +132     
  Branches     2726     2746      +20     
==========================================
+ Hits        19116    19240     +124     
- Misses       7065     7066       +1     
- Partials      933      940       +7     

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

[Copilot]

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 4 comments.

[Copilot]

Pull request overview

Copilot reviewed 8 out of 8 changed files in this pull request and generated 2 comments.

[Copilot]

Using flatten() creates a copy of the array. For better performance with potentially large step arrays, use steps_np.ravel() which returns a view when possible, or iterate directly with steps_np.flat which is a flat iterator.

Suggested change

	if not all(s == 1 for s in steps_np.flatten()):
	if not all(s == 1 for s in steps_np.flat):