Bazel configuration missing for Python bindings?

Nobody has claimed this yet.

I'm interested in using the torch-mlir python bindings in a project that uses Bazel as its build system. Based on the [developer docs](https://github.com/llvm/torch-mlir/blob/main/docs/development.md#bazel-build), it seems like the Bazel configs for torch-mlir are mostly focused on supporting `torch-mlir-opt`, and indeed I don't see any CAPI or Python binding rules in the main [Bazel build file](https://github.com/llvm/torch-mlir/blob/main/utils/bazel/torch-mlir-overlay/BUILD.bazel).

I think what's needed to do this is:
* The CAPI library (from `torch-mlir-c/Dialects.h`). This can be built with `mlir_c_api_cc_library` from `@llvm-project//mlir:build_defs.bzl`.
* The pybind11 extension library (from `python/TorchMLIRModule.cpp`).
* A tablegen rule to generate the Python wrapper code for ops (`_torch_ops_gen.py`).
* (Maybe) A convenience `py_library` that bundles the above and any MLIR Python dependencies.

There are examples of these in the MLIR project's [Bazel file](https://github.com/llvm/llvm-project/blob/main/utils/bazel/llvm-project-overlay/mlir/BUILD.bazel).

Is there interest in this from the maintainers? I'd be willing to contribute these changes if so.

Start with docs/development.md#bazel-build and utils/bazel/torch-mlir-overlay/BUILD.bazel to understand the existing Bazel setup. Compare the requested CAPI, pybind11, and TableGen rules with the examples in the MLIR project's BUILD.bazel. Done means the torch-mlir Python bindings and their dependencies can be built through Bazel.

Written by the indexing model from the issue text.