Automatic experiment monitor#521
Conversation
|
After talking with Chris, will update this to show the position in the queue of the current job and the total number of jobs. |
chriseclectic
left a comment
chriseclectic
left a comment
There was a problem hiding this comment.
This PR needs some refactoring after #599 which changed some of these methods significantly. Now there is separate methods for job_status and analysis_status which you could use to get the status for each part separate (and status is combined one for full experiment). For the analysis monitoring part you could use the _analysis_callbacks ordered dict to find the first AnalysisCallback.status that is currently running to get the i/N part for analysis.
I also wonder if we should try and use a package like https://github.com/tqdm/tqdm
to make a progress bar for the monitor.
| The output is in format "Experiment Status:" followed by one of the possible | ||
| * JOB (1/N) `STATUS` if the experiment is in the job phase. For full possible statuses of | ||
| each job, see the :meth:`status` method. | ||
| * ANALYSIS IN PROGRESS if the experiment is in the analysis phase. |
There was a problem hiding this comment.
After #599 this could be changed to something like Analysis (1/N) STATUS` like for jobs, since you should be able to see how many analysis jobs there are and get the status analysis status.
|
|
3 participants
Summary
Adds
monitor()to experiment data, which displays the current status fromstatus(). Run by callingexp_data.monitor().