Builds on top of rapidsai#1769 

- [x] Removes `libraft-ann-bench` C++ based package
- [x] Creates `raft-ann-bench` packages that includes C++ tests as well as Python scripts
    - [x] `raft-ann-bench` package includes all tests for CPU and GPU
    - [x] `raft-ann-bench-cpu` package that does not depend on CUDA or RAFT GPU code
- [x] Update docs
- [x] Test artifacts and scripts in CI  
- [x] Minor code cleaning

Some changes include:

- Use `RAPIDS_DATASET_ROOT_DIR` env variable to indicate location of datasets (optional) consistent with other repos: https://docs.rapids.ai/maintainers/datasets/
- CPU and GPU packages are built in the existing GPU build GHA. Only the CUDA 12 jobs build the CPU packages. 
- Small change for invocation of scripts, for example: `python bench/ann/run.py --dataset deep-image-96-inner` is now `python -m raft-ann-bench.run --dataset deep-image-96-inner`, but still scripts meant to be invoked from the command line. 

Future improvements: 

- Remove use of popen python scripts from python scripts. 
- Improve printing and logging
- Allow functions of package to be called from python scripts. 

Closes rapidsai#1744

Authors:
  - Dante Gama Dessavre (https://github.com/dantegd)
  - Artem M. Chirkin (https://github.com/achirkin)
  - Divye Gala (https://github.com/divyegala)
  - Corey J. Nolet (https://github.com/cjnolet)

Approvers:
  - Corey J. Nolet (https://github.com/cjnolet)
  - Divye Gala (https://github.com/divyegala)
  - AJ Schmidt (https://github.com/ajschmidt8)
  - Artem M. Chirkin (https://github.com/achirkin)

URL: rapidsai#1773