Add masking #55
Add masking #55
Conversation
…ders for masking in pixels, wavelength and theta
| " | (x > sc.scalar(86e6, unit=\"ns\")),\n", | ||
| " # Mask in wavelength\n", | ||
| " WavelengthMask: lambda x: (x > sc.scalar(2.0, unit=\"angstrom\"))\n", | ||
| " & (x < sc.scalar(2.7, unit=\"angstrom\")),\n", |
There was a problem hiding this comment.
Where does this mask come from? Is there an artefact in the data that needs to be masked?
There was a problem hiding this comment.
No there is no artifact. I was just to illustrate things.
I probably should make a separate example, either in a new section at the bottom or in a new notebook.
| "source": [ | ||
| "## Create pipeline using Sciline\n", | ||
| "\n", | ||
| "We use the basic providers available in `essdiffraction` as well as the specialised `powder` and `geant4` providers." |
There was a problem hiding this comment.
There are no basic providers, only ones for powder and geant4.
src/ess/dream/io/nexus.py
Outdated
| - The high-resolution and SANS detectors have a very odd numbering scheme. | ||
| The scheme attempts to follows some sort of physical ordering in space (x,y,z), | ||
| but it is not possible to reshape the data into all the logical dimensions. |
There was a problem hiding this comment.
| - The high-resolution and SANS detectors have a very odd numbering scheme. | |
| The scheme attempts to follows some sort of physical ordering in space (x,y,z), | |
| but it is not possible to reshape the data into all the logical dimensions. | |
| - The high-resolution and SANS detectors have a very odd numbering scheme. | |
| The scheme attempts to follows some sort of physical ordering in space (x,y,z), | |
| but it is not possible to reshape the data into all the logical dimensions. |
To get proper formatting. (Note the blank line)
|
@jl-wynen for now, the notebooks only work when computing d-spacing from positions, and not from a calibration file. |
|
I know an approach but it may not be good. For CIF files we need to write data in tof. So the idea was to do the reduction in d-spacing and then convert back to tof using fake positions. You could use a similar approach here where you essentially determine calibrated positions by converting back from d-spacing. There is some ambiguity here about how you define positions and tof, though. So you / we would have to make a judgement call. But since we need this anyway for CIF, we can use the same mechanism here. |
Co-authored-by: Jan-Lukas Wynen <jan-lukas.wynen@ess.eu>
src/ess/dream/io/nexus.py
Outdated
| from ess.reduce import nexus | ||
|
|
||
| DETECTOR_BANK_RESHAPING = { | ||
| "endcap_backward_detector": lambda x: x.fold( |
There was a problem hiding this comment.
Make this into sizes, not lambdas
| ) | ||
| from .types import DetectorInfo | ||
|
|
||
| DETECTOR_BANK_SHAPES = {"powgen_detector": {"bank": 23, "column": 154, "row": 7}} |
src/ess/powder/grouping.py
Outdated
| def focus_data( | ||
| data: MaskedData[RunType], | ||
| dspacing_bins: DspacingBins, | ||
| twotheta_bins: Optional[TwoThetaBins] = None, |
There was a problem hiding this comment.
Maybe request a new data type to get data binned in theta, as in sans Qxy?
| " TofMask: lambda x: (x < sc.scalar(0.0, unit=\"us\"))\n", | ||
| " | (x > sc.scalar(16666.67, unit=\"us\")),\n", | ||
| " # Mask in wavelength\n", | ||
| " WavelengthMask: lambda x: (x > sc.scalar(0.18, unit=\"angstrom\"))\n", |
src/ess/dream/data.py
Outdated
| """Data for tests and documentation with DREAM.""" | ||
|
|
||
| from pathlib import Path | ||
| # from pathlib import Path |
…ead of using Optional[TwoThetaBins]
This adds the framework for masking according to various coordinates (
pixel_id,tof,wavelength,two_theta).It also attempts to standardize the Powgen and dream-geant4 use cases, in terms of loading files and keeping logical dimensions.
This is part of #50, but it only implements the reduction workflow side of things, and not an interactive masking tool, which is defered to a later PR.
(Apologies about the noise with quotation marks, I accidentally ran
blackat some point on the files, after the config forblackwas removed due to the switch toruff, and it therefore did not ignore quotation marks)