ponder-lab / Hybridize-Functions-Refactoring / 27298743008

79%

push

github

Recognize `tf.Tensor` subtypes in the type-hint check (#599)

* Recognize `tf.Tensor` subtypes in the type-hint check

`Parameter.hasTensorTypeHint` matched only the exact `tf.Tensor` FQN, so a
parameter annotated with a tensor-equivalent subtype (`tf.RaggedTensor`,
`tf.SparseTensor`, `tf.Variable`, `tf.IndexedSlices`) was not classified as
tensor-typed via the Phase 1 type-hint path. Recognize those subtype FQNs via a
`TF_TENSOR_TYPE_HINT_FQNS` set. The `*Spec` descriptors (`TensorSpec`,
`RaggedTensorSpec`, …) are deliberately excluded — they describe tensors but are
not tensors.

Add positive tests for the three subtypes plus a negative test pinning that
`tf.RaggedTensorSpec` is not classified.

Closes #434

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* Extract named constants for the tensor-subtype FQNs

Replace the inline subtype FQN string literals in `TF_TENSOR_TYPE_HINT_FQNS`
with named constants (`TF_RAGGED_TENSOR_FQN`, `TF_SPARSE_TENSOR_FQN`,
`TF_VARIABLE_FQN`, `TF_INDEXED_SLICES_FQN`), each documented like the existing
`TF_TENSOR_FQN`. Pure refactor; behavior unchanged.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>

3 of 3 new or added lines in 1 file covered. (100.0%)

1389 of 1781 relevant lines covered (77.99%)

0.78 hits per line