Cirq
Cirq copied to clipboard
Make it easier to request retry if quantum engine job id is not unique
If you accidentally generate a not-random-enough job id, the engine will fail with a generic EngineError.
Possibly make it easier to manually handle this error or intelligently retry with a modified ID
If you accidentally generate a not-random-enough job id, the engine will fail with a generic EngineError.
Do you mean that you send a job with a duplicate id? Or is there a specific check for level of randomness?
When you submit an engine job, you must provide a job id. By default, cirq will generate a nice long random string for it. I was trying to give more "meaningful" job ids so I could see what was going on by scanning the web interface. Easier tagging / labeling of jobs would probably fix the root problem.
But yes, if you [accidentally?] send a job with a duplicate id it doesn't give a particularly helpful error message or type.
Thanks, that makes sense. This can be only fixed by Googlers for now, until there is an open access.
@wcourtney Is this still a valid issue? Can we close it?