FindClusterFaces Algorithm

[Owen Arnold]

This algorithm is a key component of #9375, which will be used to display cluster workspaces in the SliceViewer.

This algorithm inspects an image (MDHistoWorkspace) containing clusters, and takes an optional PeaksWorkspace as a filter. It then builds a table of all the edges of each cluster. The workings have been outline in #9375.

I have considered using MarchingCubes for this type of analysis, but the interpolation to generate polygons will not give the true region of the cluster.

[Owen Arnold]

This algorithm may be very useful in its own right.

[Owen Arnold]

refs #9428. Basic algorithm with tests.

Changeset: 25a5ca6cbf16fe88a3082a70f556215e26349636

[Owen Arnold]

refs #9428. Working on adding face-touching neighbour finding.

Changeset: c7d22a8267527bf12a504c1a9ee2d3dd0a6dfc4a

[Owen Arnold]

refs #9428. Extend functional and performance tests.

Changeset: 6a182aab4d1455c34421e5e1993ebf3e541dc30e

[Owen Arnold]

refs #9428. Extend interface to include new find signature.

Changeset: 3ffa9c4b5e54a55768bac804c4314a638e8549f2

[Owen Arnold]

refs #9428. Use face neighbour finding and test.

Changeset: 20f239691b8fc1c27939ac703834dac7a040872f

[Owen Arnold]

refs #9428. New utility PeakClusterProjection.

Changeset: 81804fe811becbcdaa729e68dee51f703594bf9d

[Owen Arnold]

refs #9428. Extended tests for PeakClusterProjection.

Changeset: c7172bd32477ec54551e27e6a8b0270fda4f7786

[Owen Arnold]

refs #9428. Fix errors and tests.

Error in IntegratePeaksUsingClusters test introduced (previously) and now fixed. Windows disambiguation of exact enum to use.

Changeset: 4a5a430bdc62ee714ba3d44cd3373abf9a551ef9

[Owen Arnold]

refs #9428. Swap in PeakClusterProjection into integration.

refactoring step. Common code usage.

Changeset: 88d62a6e6fea114c46b220a344be174adac606ce

[Owen Arnold]

refs #9428. More tests and performance tests.

Increase test coverage. Add performance tests, and refactor the implementation a little.

Changeset: c38ea5a71eb69e14c8f36a36f4a0a0c1328789f2

[Owen Arnold]

refs #9428. Fix type conversion warnings.

Changeset: 7e27b82d8ef9623bb266172d0ce11dbd556d5593

[Owen Arnold]

refs #9428. Fix SignalRangeTest.h Iterator usage.

Changeset: 9ad566c3e909e6b593e186adb3f0d3385f426dd4

[Owen Arnold]

refs #9428. Fix cmakelists.txt for test.

Changeset: b66546a2675e11cdc00f8cbb955803aece9f9dd3

[Owen Arnold]

refs #9428. Fix doxygen warning.

Changeset: 6b3db62ff92705106263b910e554c04e082b64e9

[Owen Arnold]

refs #9428. Fix type comparison warning.

Changeset: 934fb6c9279bf5516e7303068f1978fc2259bc87

[Owen Arnold]

refs #9428. Fix type doxygen warning.

Changeset: d2efdaf8a654103a504d27b30c144b2cbb88cbb5

[Owen Arnold]

Tester:

1) There are extensive unit tests covering the workings of this new algorithm.

2) Running the following python script will also demonstrate this working.

ws_name = "TOPAZ_3132"
filename = ws_name +"_event.nxs"
ws = LoadEventNexus(Filename=filename,FilterByTofMin=3000, FilterByTofMax=16000)

# Convert to Q space
LabQ = ConvertToDiffractionMDWorkspace(InputWorkspace=ws, LorentzCorrection='0',
        OutputDimensions='Q (lab frame)', SplitInto=2, SplitThreshold=150)
        
# Find peaks
PeaksLattice = FindPeaksMD(InputWorkspace=LabQ,MaxPeaks=100)

        
# Find the UB matrix using the peaks and known lattice parameters
FindUBUsingLatticeParameters(PeaksWorkspace=PeaksLattice, a=10.3522, b=6.0768, c=4.7276,
                alpha=90, beta=90, gamma=90, NumInitial=20, Tolerance=0.12)
                
# And index to HKL            
IndexPeaks(PeaksWorkspace=PeaksLattice, Tolerance=0.12)

HistoMDQLab = BinMD(InputWorkspace=LabQ,AlignedDim0='Q_lab_x, 2.7, 3, 50',AlignedDim1='Q_lab_y, -0.9, -0.6, 50',AlignedDim2='Q_lab_z, 9.2, 9.5,  50')   
PeaksLattice_Integrated_Clusters, ClusterImage = IntegratePeaksUsingClusters(InputWorkspace=HistoMDQLab, PeaksWorkspace=PeaksLattice, Threshold=100000)

# Show Cluster with Peak Overlayed. Only One point where a PeaksWorkspace intersects with a Cluster!
svw = plotSlice(ClusterImage)
sv = svw.getSlicer()
pp_all = sv.setPeaksWorkspaces([PeaksLattice_Integrated_Clusters])
pp = pp_all.getPeaksPresenter(PeaksLattice_Integrated_Clusters)
pp.zoomToPeak(30)
svw.zoomBy(0.03)

# So finding with the PeaksWorkspace as a filter should yield results that have the SAME value for 'ClusterId 'in all rows!
with_filter = FindClusterFaces(InputWOrkspace=ClusterImage, FilterWorkspace=PeaksLattice)
# No such restriction when filtering is not applied
no_filter = FindClusterFaces(InputWOrkspace=ClusterImage)

testRow = with_filter.row(0)
clusterId = testRow['ClusterId']
filtered_count = with_filter.rowCount()
unfiltered_count = 0
for i in range(no_filter.rowCount()):
        _row =  no_filter.row(i)
        if _row['ClusterId'] == clusterId:
                unfiltered_count += 1
print 'Target ClusterId occurances in Filtered Output', filtered_count
print 'Target ClusterId occurances in UnFiltered Output', unfiltered_count

• Look at the OutputWorkspace called 'with_filter'. All entries for ClusterId should be the same!
• Look at the OutputWorkspace called 'no_filter'. Not all entries should be the same, but 'with_filter' should be a subset of 'no_filter'

[Samuel Jackson]

Merge remote-tracking branch 'origin/feature/9428_find_cluster_edges'

Full changeset: ​33b2e4f5b4ba0321166549c1f373854c6bb3b7bb

[Owen Arnold]

refs #9428. Extend solution to scaling RHSWS

Changeset: 64c42fb7b13200faae409ea48c1835cd4947d6c0

[Owen Arnold]

refs #9428. Extend tests for multithreaded zero search

Changeset: c14d80f835abfff4d771cb2e519938fe8c696489

[Owen Arnold]

refs #9428. Protect against bad scale factors.

Changeset: 8d356854f77dbf48a635437468d6de68fa5e4e93

[Owen Arnold]

refs #9428. Add documentation and tests.

Incomplete functionality. Had to fix polynomial usage over Max's implementation in quick.

Changeset: 532039904a83d2fc6633c5e9efcadf773adfeddc

[Owen Arnold]

refs #9428. Remove need for Params in workflow algs

Changeset: d6d43dabc3ea2e9a0a8c8b3c5a56451d1d4cd284

[Owen Arnold]

refs #9428. Stitch as part of process.

If stitch is checked, then the stitched output workspace will be generated as part of the processing. If it is done after, then it is completed as part of the plot process

Changeset: 1c2277a435ba588be8fd13282ded508b3c08bccc

[Stuart Campbell]

This ticket has been transferred to github issue ​10271