MantidEV misses Monoclinic C for TOPAZ 7424

[Dennis Mikkelson]

One of the possible cell types for the sample in run 7424 on TOPAZ should be Monoclinic C. This is missing from the list of possible cells in MantidEV.

[Dennis Mikkelson]

RE 7244_MantidEV_misses_Monoclinic_C

Replaced ScalarUtils::GetSignRelatedUBs() with ScalarUtils::GetRelatedUBs(). The new method not only tries negating pairs of edges, but also forms a set of permutations of the edges that both preserve handedness and could still form Niggli cells to within experimental error. With this fix MantidEV now finds the Monoclinc C cell, so this fixes the problem. The test code still needs to be updated for this new method.

refs #7244

Changeset: 79ec229ece7a39c78a2e916a23ab96514903b7e0

[Dennis Mikkelson]

RE 7244_MantidEV_misses_Monoclinic_C_for_TOPAZ_7424

Updated tests and added test for the Monoclinic C case for run TOPAZ 7424.

refs #7244

Changeset: f02a01df3c4f53774b79a7b3dd19f93b80208242

[Dennis Mikkelson]

NOTES TO TESTER:

The branch name is: feature/7244_MantidEV_misses_Monoclinic_C_for_TOPAZ_7424

Testing this will require processing the event data from run TOPAZ_7424, since the problem is only apparent in certain fairly rare test cases. To verify that the problem exists before this fix, carry out the following steps with Mantid BEFORE checking out and building the new code.

1. Start MantidEV (SCD Event Data Reduction) from the Interfaces menu.

2. On the Select Data tab, use the Browse button for loading an event file to choose TOPAZ_7424_event.nxs. Set the "Max |Q| to Map to MD" to 15, and mark "Apply Lorentz Correction" as true. Press Apply to load the data.

3. On the Find Peaks tab select "Find Peaks" and specify "Estimated Max of abc" to be 10, "Number of Peaks to Find" to be 120 and "Min Intensity(above ave) to be 100. Press Apply to find the peaks.

4. On the Find UB tab, select "Find UB Using FFT" and specify "Estimated Lower Bound for a,b,c" to be 4. "Estimated Upper Bound for a,b,c" to be 8 "Tolerance" to be 0.12 Press Apply to actually find the UB matrix.

5. On the Choose Cell tab, select "Show Possible Cells" and set the "Max Scalar Error" to 0.2. Pressing Apply will show only Triclinic P as a possible cell. Raising the "Max Scalar Error" to 0.8 and pressing apply will include a Monoclinic I cell, with a relatively large scalar error around 0.7, but still no Monoclinic C.

6. Checkout and build the branch with the fix.

7. Repeat steps 2-5. This time the list of possible cells will include Monoclinic C with a small scalar error around 0.0068.

[Vickie Lynch]

Works correctly for TOPAZ_7424

[Stuart Campbell]

This ticket has been transferred to github issue ​8090