Checks for no valid requests in processed dataset

[jaredoconnell]

Summary

Adds warnings and errors so that problems in the datasets are visible, rather than silently failing.
In main, if the dataset isn't processed correctly, GuideLLM deadlocks while the datasets are pending.

Details

• Adds failure condition in scheduler/worker_group.py to abort if there are no requests.
• Adds some other warnings when there are no matching columns

Test Plan

Currently, due to the other bugs, this command triggers the problem:

guidellm benchmark run \
  --target http://localhost:8000/ \
  --data "abisee/cnn_dailymail" \
  --data-args '{"name": "3.0.0"}' \
  --data-column-mapper '{"text_column":"article"}'

• "I certify that all code in this PR is my own, except as noted below."

Use of AI

• Includes AI-assisted code completion
• Includes code generated by an AI application
• Includes AI-generated tests (NOTE: AI written tests should have a docstring that includes ## WRITTEN BY AI ##)

[sjmonson]

See my new comments above.

[sjmonson]

My concerns are resolved. Leaving the comments open that @dbutenhof added to in-case any of those still need to be addressed.