bk
Details
Details
- Reviewers
- None
Diff Detail
Diff Detail
- Repository
- R1 dagster
- Branch
- mappable_pipeline
- Lint
Lint Passed - Unit
No Test Coverage
Event Timeline
Comment Actions
no show stoppers i can see - figuring out a clean way to do what is currently handled by deconstruct_step_key will be an interesting challenge
| python_modules/dagster/dagster/core/definitions/output.py | ||
|---|---|---|
| 99–103 | nit: is_mappable not a constructor arg but just result of overridden method | |
| python_modules/dagster/dagster/core/events/__init__.py | ||
| 203–204 ↗ | (On Diff #25124) | would expect this to go be in event_specific_data |
| python_modules/dagster/dagster/core/execution/memoization.py | ||
| 63–67 ↗ | (On Diff #25124) | one value of having separate types is failing check.inst calls when mappable things show up where they are not expected |
| python_modules/dagster/dagster/core/execution/plan/objects.py | ||
| 163–233 | think there is some value in modeling the unresolved input more clearly as a separate thing - this class is already a bit confusing and having the mappable / resolve like this makes it even more confusing | |
| 356 | definitely weary of using inheritance here | |
| python_modules/dagster/dagster/core/execution/plan/plan.py | ||
| 307–326 | will have to come up with a clean way to model this all and update callsites accordingly | |