Move BUILD_NAMEDTENSOR macro out of header areas

[zou3519]

Stack from ghstack:

• Move BUILD_NAMEDTENSOR macro out of header areas #25256 Move BUILD_NAMEDTENSOR macro out of header areas
• Fix dependency by moving Dimname.{h,cpp} NamedTensor.{h,cpp} to core/ #25280 Fix dependency by moving Dimname.{h,cpp} NamedTensor.{h,cpp} to core/
• Include the correct header for make_unique in named tensor headers #25178 Include the correct header for make_unique in named tensor headers
• Implement name inference for torch.matmul #25177 Implement name inference for torch.matmul
• Implement name inference for torch.bmm #25123 Implement name inference for torch.bmm

Context: I am starting to work on removing the BUILD_NAMEDTENSOR flag.
Here is the approach:

• Move the macro out of header areas
• Include a new enable_namedtensor.h header that does a #ifndef BUILD_NAMEDTENSOR #define BUILD_NAMEDTENSOR.
• Include enable_namedtensor.h where necessary. This only really needs
  to happen in two files (c10/TensorImpl.h, ATen/Dimname.h).
• Incrementally delete usages of the BUILD_NAMEDTENSOR macro later.

The alternative is to straight up delete all instances of
BUILD_NAMEDTENSOR. This alternative could be disruptive, lead to merge
conflicts, and isn't incremental.

Along with the above, some work needs to be done on feature flagging
named tensors, and merging the namedtensor CI with the regular CI, and
communicating with devs. This work will too be done incrementally.

Test Plan

• [namedtensor ci]

gh-metadata: pytorch pytorch 25256 gh/zou3519/121/head

Differential Revision: D17076893