Remove non-generic nexus components #112
Conversation
src/ess/reduce/nexus/types.py
Outdated
| def to_snx_selection(self, *, for_events: bool) -> snx.typing.ScippIndex: | ||
| if self.value == slice(None, None): | ||
| return () | ||
| if isinstance(self.value, slice): | ||
| if for_events: | ||
| return {'event_time_zero': self.value} | ||
| return {'time': self.value} | ||
| return self.value |
There was a problem hiding this comment.
Duplicate of _nexus_loader._to_snx_selection. Why did you change the loader to take a raw selection instead of a PulseSelection object?
There was a problem hiding this comment.
Because PulseSelection is now generic, so that didn't work.
Thought I removed this (I was moving it back and forth), good catch.
| # include the end | ||
| bounds = sc.arange( | ||
| 'chunk', info.start_time, info.end_time + chunk_length, chunk_length | ||
| def GenericNeXusWorkflow() -> sciline.Pipeline: |
There was a problem hiding this comment.
| def GenericNeXusWorkflow() -> sciline.Pipeline: | |
| def LoadNeXusWorkflow() -> sciline.Pipeline: |
Since you removed 'generic' everywhere else.
There was a problem hiding this comment.
"Generic" here means "cross-technique".
There was a problem hiding this comment.
You didn't use that word for the other workflows in this file. And they are also cross-technique
There was a problem hiding this comment.
I don't see the point in changing naming in this context. If you think it is really important we can sit down and have a session on overall naming conventions.
There was a problem hiding this comment.
I don't think it is very important. I just think that 'GenericNeXusWorkflow' sounds like it does something conceptually different than 'LoadMonitorWorkflow'. Where actually, it just loads data and is as generic as 'LoadMonitorWorkflow'.
These were added initially when it was not clear that we would go with a fully generic workflow, and left in since I thought they might be useful for debugging or general exploratory work. However, it now proves that they simply add friction in further code changes, so there is no benefit.
Fixes #103 by removing it.
I have checked on ESSsans that downstream packages will simply have to update a few imports, i.e., updating this will be trivial.
There are no code changes here, apart form removing forwarding and updating type-hints.