Better naming for github action runs

[sujik18]

Fix for #134

[github-actions]

MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅

[arjunsuresh]

In the line above should we replace "build" with something better? May be "rgat-inference-run" ?

[arjunsuresh]

@sujik18 I was referring to the previous line where "build" is used. Do you know if the "name" field can override the job name "build"? If so, then it is fine.

[sujik18]

@arjunsuresh Yes, I believe that by mentioning name field it will override the "build" as it did work for other tests.

[arjunsuresh]

If you see the action here it still shows "build": https://github.com/mlcommons/mlperf-automations/actions/runs/13100736773

[sujik18]

@arjunsuresh I tried replacing the build value with another job key value, but the test name is still showing as build. You can see it here https://github.com/mlcommons/mlperf-automations/actions/runs/13104065296/workflow
I double-checked for any syntax errors, but there aren't any. Even if it was a GitHub internal issue, it would have also affected other test names, but that isn't the case, so I'm not sure why it's happening.

[arjunsuresh]

@sujik18 if the pull_request_target is specified in the github action, then the action is running from the base branch and so the PR changes won't be reflected there. But you can see for the changes in the github actions with "on:pull_request".

[arjunsuresh]

@sujik18 the change is working fine 👍. Can you please revert back to pull_request_target - once the PR is merged, from next PR onwards the tests will run as expected.

[sujik18]

@arjunsuresh I now understand that I was only checking the on:pull_request_target event. The recent commit was unnecessary. I will revert back to pull_request_target, and we can simply ignore the last commit to proceed with merging the PR

[arjunsuresh]

Sure. The changes are fine for the github actions. Have you seen my comment on the logging part in the github issue?

[sujik18]

Yes, actually by mistake those commits got included with this PR instead of a new one, wrt to logger issue, apart from making the induvial script to utilize global logger, do we also need each script to write the logs in a single mlc-log.txt file,
I guess that would be a better approach for maintaining the logs?
Currently each script file is generating its own log.txt in its own directory. For that we need to change the default value given to the path for setup_logging function call, as the folder directory of mlc seems to have been changed, the current default value seems to be invalid.