Use pixel masks from XML file for Loki #96
Conversation
| @@ -87,46 +106,94 @@ def _merge_runs(data_groups: Iterable[sc.DataGroup], entries: Iterable[str]): | |||
| return out | |||
|
|
|||
|
|
|||
| def merge_sample_runs( | |||
There was a problem hiding this comment.
I had to specialise the types here for SampleRun and BackgroundRun, instead of having a single function with ScatteringRunType, because if not there would be multiple providers for LoadedFileContents[RunType] as the to_file_contents below also matches.
I don't think it's pretty. We also need to manually insert the merge_sample_runs and merge_background_runs into the pipeline if we want to use a parameter table with multiple runs.
Is there a better way to achieve this?
I basically want to merge the runs if a parameter table is supplied, but do nothing otherwise.
There was a problem hiding this comment.
We can ignore this problem for now, as we plan to have a more generic solution.
| @@ -87,46 +106,94 @@ def _merge_runs(data_groups: Iterable[sc.DataGroup], entries: Iterable[str]): | |||
| return out | |||
|
|
|||
|
|
|||
| def merge_sample_runs( | |||
There was a problem hiding this comment.
We can ignore this problem for now, as we plan to have a more generic solution.
src/esssans/loki/masking.py
Outdated
| lowcounts_straw_mask: Optional[DetectorLowCountsStrawMask], | ||
| beam_stop_mask: Optional[DetectorBeamStopMask], | ||
| tube_edge_mask: Optional[DetectorTubeEdgeMask], | ||
| def apply_pixel_masks( |
There was a problem hiding this comment.
Why is this Loki-specific?
There was a problem hiding this comment.
Yes, this is another slightly annoying thing I had to work around.
In isis, the masks are applied on RawDataWithComponentUserOffsets, while in loki, they are applied on RawData.
Can we do it differently? I am not sure we want to introduce the ComponentUserOffsets everywhere?
I could make the apply_pixel_masks a common function and then have wrappers inside isis and loki that map the correct types, but I feel that would not be ideal either?
There was a problem hiding this comment.
What about Skadi etc.? My question was not necessarily if it can be the same for the ISIS instruments, more about all ESS instruments.
There was a problem hiding this comment.
Ok, will move into a common place for loki and leave the isis one in the submodule 👍
Replaces #91
We had an issue with masking when merging multiple runs in the Loki workflow,
For example, to mask the straws at the top and bottom of the panel, we summed the counts inside each straw, and compared that to the number of pixels in each straw. This means that all you need to do is merge the events from 2 runs, and you will have approximately double the events but the number of pixels stays the same.
So now, in some straws, double the events means you are above the threshold and you start to mask less straws. The more runs you merge, the less straws you mask.
Here, we load an XML file containing the detectors to mask, as is done in the Zoom workflow.
Other changes:
poochfiles by configuringDataFolderload_nexus(instead of a file list). If we want to merge runs, we now have to set up a parameter table (added a section in the loki notebook that shows how to do this).Note that the changes in masks seem to result in slightly worse results at low Q:
Before:
After:
This may suggest that we need to mask more pixels close to the beam stop? But that would be for the instrument scientist to decide.