Framework Changes

Algorithms

New features

• (#40367) UnaryOperation algorithms like ReplaceSpecialValues, Logarithm now support ragged workspaces.

• Several algorithms have had flowcharts added to document their processing steps:

  • (#40063) LoadWANDSCD v1

  • (#40120) ConvertWANDSCDtoQ v1

  • (#40234) ConvertHFIRSCDtoMDE v1

• AlignAndFocusPowderSlim v1 has some new parameters:

  • (#40067) OutputSpectrumNumber, which allows a single bank of data to be loaded.

  • (#40122) ProcessBankSplitTask, which can split its output into multiple workspaces when set to True.

  • (#40377) UseFullTime, which can split events based on full time (pulse time+TOF) when True

  • (#40669) ReadSizeFromDisk, EventsPerThread, and LogBlockList are existing parameters with new defaults to provide faster execution.

• (#40493) AlignAndFocusPowderSlim v1 will now set the uncertainty values on the output workspace.

• (#40107) HFIRPowderReduction v1 is a new algorithm, which is currently just the UI elements for loading and reducing data.

• (#40128) CreateDetectorTable v1 now supports the option PickOneDetectorID. This returns a column of detector IDs as a column of integers containing only one detector ID per row. By default this option is disabled, so by default the column of detector IDs will consist of strings of one or more detector ID per row. This new option was introduced to be used in the background of the new instrument view interface, but might also be useful if the user wants a single detector ID per row.

• (#40227) CreateGroupingByComponent v1 is a new algorithm that allows for the creation of GroupingWorkspaces, where the groups are formed by combining components under a given parent component. Components which should be combined are identified by providing strings to indicate what their names should/should not contain. Each of these groups can be optionally subdivided into a further, GroupSubdivision, number of groups.

• (#40327) ReorientUnitCell v1 is a new algorithm that allows the selection of the unit cell most aligned to goniometer axes.

• (#40332) AddLogSmoothed v1 is also new, and allows for the addition of smoothed time-series log data.

• (#40450) AddLogInterpolated v1 is new, and provides a method for interpolating time series sample logs.

• (#40420) LoadNGEM v1 has been extended to allow for data to be loaded directly into histograms, by setting the PreserveEvents property to False. For workflows where preserving events is not necessary, this significantly reduces load time and memory usage.

• (#40429) MaskDetectorsIf v1 now allows the user to input the start and end index of the range to apply the mask on.

• (#40656) Load v1 and the file loader search process now use a lazy file descriptor to check for the correct loading algorithm, further reducing the time taken to load files.

• (#40444) ApplyDetailedBalanceMD v1 has a new parameter, RescaleToTemperature, which allows the scaling of data from one temperature to another.

Bugfixes

• (#40029) mantid.api.MatrixWorkspace.getNumberBins() has been replaced by mantid.api.MatrixWorkspace.blocksize(), as the former is deprecated.

• (#40606) GenerateEventsFilter v1 will now be able to handle cases where there is no second pulse in the ROI.

• (#40336) CreateDetectorTable v1 will now return the workspace index as an integer, rather than a floating-point number.

Deprecated

• (#40380) GetTimeSeriesLogInformation v1 has been deprecated. Use workspace.run().getStatistics("log_name") instead to access time series log statistics directly.

• (#40625) LoadMcStasNexus v1 has been deprecated. Use LoadMcStas v1 instead - it is more up-to-date.

Fit Functions

Bugfixes

• (#40267) Gaussian has improved stability of fits by changing the internal ‘active’ parameter to 1/sigma instead of 1/sigma^2.

• (#40267) The normalisation in histogram evaluation of Gaussian function and Jacobian has been fixed such that they are now consistent with CentrePoint evaluation for small bin-widths.

Data Objects

New features

• (#40054) Table Workspaces now support mantid.api.ITableWorkspace.columnArray() for retrieving a column as a NumPy array. For example, with import numpy as np: table_ws.columnArray("Index") returns np.array([0, 1, 2, 3]).

Bugfixes

• (#40551) MatrixWorkspace.getMemorySize() now reports the correct value for a Ragged Workspace.

Python

Bugfixes

• (#40239) SampleShapePlot (from mantidqt/plotting/sample_shape) no longer crashes upon being called from a script in workbench. Previously, this would crash as the plotting methods expected to be on the user interface thread.

• (#40545) jemalloc has been removed from the conda activation scripts because it was causing crashes with unrelated applications on Linux due to LD_PRELOAD being set globally. Behaviour is now reverted to Mantid v6.13 where jemalloc is used only when launching mantidworkbench.

Dependencies

Bugfixes

• (#40246) Pin occt to 7.9.2 to avoid conda/rattler version number parsing issue with 7.9.1 (See here).

Kernel

New features

• (#40343) Created function strmakef, which allows developers to create std::string objects using C-style formatting strings.

• (#40487) Property and IPropertyManager now support the application of multiple IPropertySettings instances to a single property instance. From Python, a property’s getSettings method will now return a vector of IPropertySetting when appropriate.

MantidWorkbench

See Mantid Workbench Changes.

Release 6.15.0