Poll Based Waiting for Job Completion #670
Conversation
MattToast
added
type: feature
MattToast
force-pushed
the
wait-for-job-end
branch
from
b46345d to
ae86890
Compare
`Experiment` was given a `wait` method that takes a collection of Launched Job IDs and will wait until the launch reaches a terminal state by either completing or erroring out. Implements a polling based solution.
MattToast
force-pushed
the
wait-for-job-end
branch
from
ae86890 to
7c633a6
Compare
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## smartsim-refactor #670 +/- ##
=====================================================
+ Coverage 40.45% 43.28% +2.82%
=====================================================
Files 110 110
Lines 7326 7053 -273
=====================================================
+ Hits 2964 3053 +89
+ Misses 4362 4000 -362
|
| :param timeout: The minimum amount of time to spend polling all jobs to | ||
| reach one of the supplied statuses. If not supplied or `None`, the | ||
| experiment will poll indefinitely. | ||
| :param interval: The minimum time between polling launchers. |
There was a problem hiding this comment.
Is param interval just for us to use for testing? Seems like user cannot define
There was a problem hiding this comment.
Exactly! It was something that, right now, could be hard coded, but if in the future we wanted to make it variable we can change the parameter. Totally willing to remove if we think the excess complexity is unnecessary in a YAGNI way!
There was a problem hiding this comment.
I wouldn't mind leaving it in and keeping it in the docs
There was a problem hiding this comment.
I say keep if but not in docstring
| :param timeout: The minimum amount of time to spend polling all jobs to | ||
| reach one of the supplied statuses. If not supplied or `None`, the | ||
| experiment will poll indefinitely. | ||
| :param interval: The minimum time between polling launchers. |
There was a problem hiding this comment.
I wouldn't mind leaving it in and keeping it in the docs
Experimentwas given awaitmethod that takes a collection of Launched Job IDs and will wait until the launch reaches a terminal state by either completing or erroring out. Implements a polling based solution.