Transformation-pattern parsing

This module parses and proves canonical round trips for rearrange, repeat, and reduce patterns.

def TorchLean.Tensor.Internal.Syntax.parseTransformPattern (source : String) (policy : IdentifierPolicy := IdentifierPolicy.pythonUnicode) : Result TransformPattern

Parse the shared pattern language of rearrange, repeat, and reduce.

theorem TorchLean.Tensor.Internal.Syntax.parseTransformPattern_render_eq_ok (source : String) (pattern : TransformPattern) (hParse : parseTransformPattern source = Except.ok pattern) : Except.map TransformPattern.render (parseTransformPattern pattern.render) = Except.ok pattern.render

Parsing a successful transformation pattern's canonical rendering succeeds with the same canonical syntax.

This applies uniformly to patterns used by rearrange, repeat, and reduce; operation-specific shape conditions belong to the checker rather than this syntax theorem.