WorkflowCollection for batch reduction #166
Conversation
src/ess/reflectometry/tools.py
Outdated
| zip(curves.keys(), scaling_factors, strict=True) | ||
| ) | ||
|
|
||
| if CriticalEdgeKey in wfc.keys(): |
There was a problem hiding this comment.
I'm not super happy about the implementation with the CriticalEdgeKey...
Suggestions welcome.
There was a problem hiding this comment.
Note that this was removed after the most recent refactor.
nvaytet
changed the title
src/ess/reflectometry/tools.py
Outdated
| The sciline workflow used to compute the targets for each of the runs. | ||
| runs: | ||
| The sciline parameters to be used for each run. | ||
| TODO: explain how grouping works depending on the type of `runs`. |
There was a problem hiding this comment.
@jokasimr it would be super if you could help me explain here the convention for the runs, when are event lists combined, and when they represent different angles 🙏
There was a problem hiding this comment.
runs is a sequence of dictionaries. The dictionaries contain sciline domain types and their associated values. The basic idea is that calling the helper is equivalent to, for each dictionary: 1. Copy the workflow, 2. Set the domain types and their values from the dictionary on the workflow, 3. Compute the requested targets.
But there are two additional rules:
If run[Filename[SampleRun]] is not a string or a path, but a collection of such, then it is interpreted as a list of files that should be combined.
If runs is a dictionary then the logic is applied to the .values() of the dictionary, and instead of returning a sequence of results we return a dictionary of results with the same keys as the input
nvaytet
changed the title
… hood and make its interface more like the Pipeline interface rather than a mix between Pipeline and a dict
This is an alternative to #155 .
Mostly, see #155 (comment) for context and motivation.
I got the Amor notebook to work,
but I still need to fix/implement the unit tests, hence this is a draft for now.But feedback is welcome.
I think some things turned out quite nicely in the notebook; creating some of the figures now appears to be simpler 🙂