Invalid scheduler exception

[scottilee]

🐛 Bug

When providing an invalid scheduler to the CLI it throws a KeyError. It would probably be better to catch the error and provide a user-friendly error message.

Module (check all that applies):

• torchx.spec
• torchx.component
• torchx.apps
• torchx.runtime
• torchx.cli
• torchx.schedulers
• torchx.pipelines
• torchx.aws
• torchx.examples
• other

To Reproduce

Steps to reproduce the behavior:

1. Run the torchx CLI with an invalid scheduler (e.g., torchx run -s machinelearning --dryrun

$ torchx run -s machinelearning --dryrun

Traceback (most recent call last):
  File "/usr/local/bin/torchx", line 8, in <module>
    sys.exit(main())
  File "/usr/local/lib/python3.10/dist-packages/torchx/cli/main.py", line 110, in main
    run_main(get_sub_cmds(), argv)
  File "/usr/local/lib/python3.10/dist-packages/torchx/cli/main.py", line 106, in run_main
    args.func(args)
  File "/usr/local/lib/python3.10/dist-packages/torchx/cli/cmd_run.py", line 271, in run
    self._run(runner, args)
  File "/usr/local/lib/python3.10/dist-packages/torchx/cli/cmd_run.py", line 206, in _run
    scheduler_opts = run_opts[args.scheduler]
KeyError: 'machinelearning'

Expected behavior

Instead of throwing a key error and showing a stack trace it should throw a user-friendly error message.

Environment

• torchx version (e.g. 0.1.0rc1): torchx-0.1.2
• Python version: Python 3.10.4
• OS (e.g., Linux): Linux
• How you installed torchx (conda, pip, source, docker): pip
• Docker image and tag (if using docker): ubuntu:latest
• Git commit (if installed from source): N/A
• Execution environment (on-prem, AWS, GCP, Azure etc): Docker
• Any other relevant information: N/A

Additional context

I have a PR ready if this is a valid bug.

[kurman]

+1