Deprecate automatic flattening for tensor products

[alexfikl]

This adds a .flatten() method to tensor products to recursively flatten them into simplices (mostly). The automatic flattening in the constructor is marked as deprecated.

Fixes #126.

[inducer]

I'm supportive of this change. How come it's marked as draft?

[alexfikl]

I'm supportive of this change. How come it's marked as draft?

Wasn't quite sure about adding that flatten: bool = True argument and outright deleting the custom __new__ implementation. Doesn't seem to break anything though, so should at least be good for a quick review 😁

[Copilot]

Pull request overview

This PR updates ModePy’s tensor-product abstractions to stop implicitly “simplifying” tensor products while providing an explicit .flatten() API, and begins deprecating constructor-time auto-flattening to address surprising behavior reported in #126.

Changes:

• Add .flatten() methods for TensorProductSpace and TensorProductShape to recursively flatten nested tensor products (returning the single remaining base directly).
• Replace __new__-based auto-simplification with constructor-time behavior guarded by a flatten flag and DeprecationWarning when auto-flattening is applied.
• Adjust hypercube-related implementations to avoid relying on 1D hypercubes being Simplex, and update tests/type-check baselines accordingly.

Reviewed changes

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

File	Description
modepy/spaces.py	Adds flatten constructor flag, emits deprecation warnings for auto-flattening, and introduces TensorProductSpace.flatten().
modepy/shapes.py	Adds flatten constructor flag + TensorProductShape.flatten(), updates face counting, and removes 1D hypercube “becomes simplex” behavior.
modepy/test/test_tools.py	Removes assertions that depended on 1D tensor products/hypercubes collapsing to simplices.
.basedpyright/baseline.json	Removes baseline entries related to the deleted overload-based __new__ implementations.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[inducer]

Thanks!