Conversation
|
pre-commit.ci autofix |
for more information, see https://pre-commit.ci
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #738 +/- ##
==========================================
+ Coverage 98.87% 98.88% +0.01%
==========================================
Files 42 42
Lines 2669 2697 +28
==========================================
+ Hits 2639 2667 +28
Misses 30 30 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
gufe/protocols/protocoldag.py
Outdated
| else: | ||
| # Is source key stable enough? | ||
| # We probably don't want to resume if gufe stability has changed | ||
| results[unit_result.source_key] = unit_result |
There was a problem hiding this comment.
Do we need to check that the unit_results are continuous?
There was a problem hiding this comment.
we now make sure that any units downstream of a node to be re-run are also re-run.
|
note: I'm taking this PR over from @IAlibay. In gufe devs meeting, we discussed needing to handle the case where a DAG A -> B -> C has a failure in B, such that both B and C are re-run. |
6b19cb9 to
4fbb80b
Compare
59d3766 to
91301e9
Compare
91301e9 to
b94e329
Compare
|
No API break detected ✅ |
| @@ -0,0 +1,23 @@ | |||
| **Added:** | |||
|
|
|||
| * ``gufe.protocol.protocoldag.execute_DAG`` now has the ability to resume DAG execution by passing in a path for results cacheing into ``unitresults_basedir`` and setting ``keep_unitresults=True``. | |||
There was a problem hiding this comment.
I'm open to additions and edits here, I just want to make sure it's captured in our changelog.
Cache a JSON of each successful ProtocolUnitResult, allowing for
execute_DAGto resume when called over a DAG that hasn't been fully executed.Checklist
newsentryDevelopers certificate of origin