Ctrl+K

reduce

reduce#

overview#

The most important method in the snapwrap utils module is reduce which, as expected, reduces diffraction data. Its only mandatory argument is a run number. On the “happy path” where the run number corresponds to a fully calibrated instrument state, reduction will execute completely with only a run-number provided.

Thus, a minimal script to reduce the data corresponding to run 61991 would be:

import snapwrap.utils as wrap

wrap.reduce(64413)

Executing the script above will result in a full reduction of run 64413. The reduced data will appear in the mantidWorbench workspace tree and, by default, have been reduced according to 3 standard pixel grouping schemes: all, bank and column. Each workspace has a different number of reduced spectra, depending on the number of subgroups defined in the pixel grouping scheme (e.g. column has 6 spectra, corresponding to the 6 columns of detector modules in SNAP).

The reduced workspaces have standard names that all begin with the word reduced followed by the units (currently only dsp for d-spacing, or qsp for momentum transfer) the pixel grouping scheme (i.e. all,’bank,column) the run number and, finally a timestamp (containing year-month-day and hours, minutes and seconds). For example:

reduced_dsp_all_064413_2025-02-18T170940
reduced_dsp_bank_064413_2025-02-18T170940
reduced_dsp_column_064413_2025-02-18T170940

would be normal output for reducing run 64413.

Note

While a bit visually jarring, the timestamps are useful allowing comparison of successively reduced runs.

Note

If you are only interested in a particular group (e.g. bank), the mantid utility to filter workpaces names can be a neat way to clean up the workspace tree. You can also use the snapwrap utility wrap.file() to keep the tree clean.

As with any python method, additional arguments can be passed to reduce inside its parentheses with simple parameter=value syntax. The available parameters are described here

Optional arguments#

cisMode#

If True various intermediate workspaces created during reduction are retained. This can be useful for troubleshooting, but it’s a bit tricky to figure out what the many workspaces are.

Caution

Be aware that retaining intermediate workspaces can use up a lot of memory

continueNoDifcal#

This flag allows a “diagnostic” reduction of data to proceed when a diffraction calibration (specifically the .h5 file containing diffractometer constants) does not exist. The default setting for this parameter is False.

Note

SNAPRed follows the Calibration Index for the relevant instrument state to find calibration files. This is built automatically when using SNAPRed to conduct a diffraction calibration workflow.

continueNoVan#

Similarly to above, this parameter allows a diagnostic assessement of data where no vanadium normalisation calibration exists. When True, a background extracted from the reduced diffraction pattern is used as an artificial normalisation function for the data. The default setting for this parameter is False.

The algorithm that extracts the the artificial normalisation (clipPeaks) has three parameters (smoothingParameter,decreaseParameters and lss), their default values are set in the config file “defaultRedcConfig.yml”, and these can be overriden by using the YMLOverride option (see below).

emptyTrash#

If True (the default) snapwrap will delete any intermediate workspaces created by SNAPRed. Sometimes it’s useful to inspect these, which can be done by setting emptyTrash=False

keepUnfocussed#

By default set to False, this flag allows the retention of the unfocussed, as loaded dataset. x-units are by default d-spacing (this can be changed using the YMLOverride option).

pixelMaskIndex#

Currently, SNAPRed can apply a pixel mask with the following caveats:

  1. the mask must exist as a mantid mask workspace and

  2. the workspace must have the name either MaskWorkspace or MaskWorkspace_N

Masks of this type are specified by specifiying the value of N (note the workspace MaskWorkspace can be indicated with either N= 0 or 1 as the first value of N used by mantid is 2).

qsp#

Default value is False. If set True data will be reduced into workspaces in units of momentum transfer with linear binning following the value of linBin.

linbin#

Defaulting to 0.01, this sets the linear bin size in Angtrom that is applied to data when momentum transfer units have been specified (following qsp=True). Linear binning is not appropriate for data in units of d-spacing and this is not supported by snapwrap.

reduceData#

By default set to True, this allows control over whether the data are reduced. If set False, snapwrap will report relevant information on the reduction (e.g. existence and location of various calibration files) but will not actually reduce data.

save#

This controls whether the reduced data are saved to disk, along with the reduction record and any auxilliary workspaces (such as pixel masks). By default, this is set to be True and every execution of reduction will create a new time stamped folder. If set False data will be reduced and workspaces created, but these are not saved.

verbose#

By default set to False, this flag allows control of the mantid logging level and also allows detailed information regarding the reduction to be output to the messages window.

YMLOverride#

Some default settings for snapwrap are set in a yml file inside the repo. It is possible to override any value by copying this file, editing as needed and then setting the value of YMLOverride to be the path to the new .yml

Contents