Returns a bit-vector indicating for each input dimension whether the lower bound is implicit rather than explicit.

An explicit bound is one that has been specified directly, e.g. by a slicing operation, and serves as a hard constraint. An implicit bound is merely informative and does not constrain indexing or composition; a typical use of an implicit bound is to specify the last-retrieved value for a remotely-stored TensorStore that supports resizing.

Implicit bounds are primarily useful on input dimensions used by OutputIndexMethod::single_input_dimension output index maps: in that case, the PropagateBounds operation may be used to update them based on an output space Box.

If an index array-based output index map depends on a given input dimension, it is an invariant of IndexTransform that the lower and upper bounds of that input dimension are explicit.

Pre:¶

valid()