[Autotuner] WIP: CLI refactor

[luarss]

This pull request introduces significant changes to the AutoTuner CLI, enhancing its usability and modularity. The most notable updates include the addition of YAML-based configuration file support, the refactoring of argument parsing into a dedicated cli module, and updates to documentation to reflect these improvements. These changes aim to simplify user workflows, improve reproducibility, and modernize the codebase.

Enhancements to CLI Usability:

• Introduced YAML-based configuration file support for AutoTuner, allowing users to specify parameters in a structured and reusable format. This reduces the need for verbose command-line arguments. (tools/AutoTuner/cli.yaml, docs/user/CLIGuideAutotuner.md) [1] [2]
• Updated documentation to include a guide on using the new configuration-based approach and adjusted example commands to reflect the new CLI syntax. (docs/user/CLIGuideAutotuner.md, docs/user/InstructionsForAutoTuner.md) [1] [2] [3]

Codebase Refactoring:

• Refactored argument parsing logic into a new cli.py module, leveraging jsonargparse for better type validation and modularity. This replaces the previous ad-hoc argument parsing in distributed.py. (tools/AutoTuner/src/autotuner/cli.py, tools/AutoTuner/src/autotuner/distributed.py) [1] [2] [3]
• Removed redundant argument parsing code from distributed.py and updated it to use the new parse_arguments function from the cli module. (tools/AutoTuner/src/autotuner/distributed.py) [1] [2]

Dependency Updates:

• Added jsonargparse to requirements.txt to support the new argument parsing functionality. (tools/AutoTuner/requirements.txt)

Minor Adjustments:

• Updated test cases and utility functions to align with the new CLI structure. (tools/AutoTuner/test/ref_file_check.py, tools/AutoTuner/src/autotuner/utils.py) [1] [2]
• Replaced outdated example commands in distributed.py with the new syntax. (tools/AutoTuner/src/autotuner/distributed.py)

These changes collectively improve the user experience, streamline the codebase, and set the foundation for future enhancements.