rework factories, enclosure, and cross section registration

[sebastian-goeldi]

Summary

1. generic_factories — register arbitrary functions as factories

• layout.py: new KCLayout.generic_factories: dict[str, Callable[..., ProtoTKCell[Any]] | Callable[..., VKCell]], plus a @kcl.generic_factory decorator (usable bare, with name=, or as a direct call) that registers a cell-producing function and wraps it with a
  guardrail raising if the returned cell's .kcl isn't this layout.
• tests/test_generic_factories.py: registration/delegation, the foreign-kcl guardrail, custom name, and independence from factories / virtual_factories.

2. Cross-section / enclosure rework (the bulk)

• Single registry: CrossSectionModel now holds one cross_sections: dict[str, SymmetricalCrossSection | AsymmetricalCrossSection] (dropped the separate asymmetrical_cross_sections dict). One unified _register; cross-kind name uniqueness is automatic.
• Canonical naming: every entry is keyed by its geometry-derived auto_name() and, when named, its explicit name. Existence is a single get(auto_name()). Unnamed entries resolve to a structurally-identical named one; registering a name promotes an unnamed
  canonical; a second name for one signature raises.
• bbox → enclosure: bbox_sections removed as a SymmetricalCrossSection field; now a read-only property delegating to the enclosure. The enclosure's structural hash now folds in bbox_sections, so auto_name() is a complete key. Asymmetric auto_name() is now a
  SHA1 over its geometry.
• radius as metadata: excluded from auto_name(), __eq__, and __hash__. Re-registering the same profile with a different radius raises (override radius at routing time instead).
• __eq__: cross-kind comparisons return False; wrapper (CrossSection/DCrossSection/asymmetric) comparisons compare bases.
• New error: CrossSectionNamingConflictError in exceptions.py.
• symmetrical kwarg: get_cross_section / get_icross_section / get_dcross_section gained symmetrical: bool | None = None (replacing the old kind literal) — None = either kind, True = symmetric (raise otherwise), False = asymmetric (raise otherwise),
  via overloads.
• Consumers updated for the single registry: serialization in layout.py (one loop, de-aliased via set()), session_cache.py (split into two pkls on dump, load both back), utils/difftest.py, schematic.py (iterates all kinds; map type widened), kcell.py.

3. Type / lint hygiene

• decorators.py: @final on WrappedKCellFunc / WrappedVKCellFunc so isinstance narrowing is clean (fixed the schematic_driven diagnostics, removed redundant # ty: ignores).
• Factories.get: key widened to object to satisfy Mapping.get's LSP contract.
• Test/conftest updates (sym_enc factory fixture, etc.).
  • decorators.py: @Final on WrappedKCellFunc/WrappedVKCellFunc so isinstance narrowing is clean (fixed the schematic_driven diagnostics, removed redundant # ty: ignores).
  • Factories.get: key widened to object to satisfy Mapping.get's LSP contract.
  • Test/conftest updates (sym_enc factory fixture, etc.).

[codecov]

Codecov Report

❌ Patch coverage is 74.17582% with 47 lines in your changes missing coverage. Please review.
✅ Project coverage is 71.00%. Comparing base (b842d75) to head (439c6b3).

Files with missing lines	Patch %	Lines
src/kfactory/cross_section.py	80.55%	8 Missing and 6 partials ⚠️
src/kfactory/kcell.py	21.42%	11 Missing ⚠️
src/kfactory/layout.py	73.17%	6 Missing and 5 partials ⚠️
src/kfactory/enclosure.py	90.00%	2 Missing and 2 partials ⚠️
src/kfactory/session_cache.py	42.85%	2 Missing and 2 partials ⚠️
src/kfactory/utils/difftest.py	0.00%	1 Missing and 1 partial ⚠️
src/kfactory/schematic.py	66.66%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #987      +/-   ##
==========================================
+ Coverage   70.96%   71.00%   +0.03%     
==========================================
  Files          72       72              
  Lines       14423    14526     +103     
  Branches     2795     2817      +22     
==========================================
+ Hits        10236    10314      +78     
- Misses       3337     3347      +10     
- Partials      850      865      +15     

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

🚀 New features to boost your workflow:

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