SANS Trans FIT option

[Gesner Passos]

Error reported by the user:

ISIS SANS "separate" option on transmission fitting, seems to always do the default “logarithmic” fit, regardless.

e.g Try, from present cycle, SANS run 22031, trans run 22027 divided by 22024 for direct beam, with mask file attached. Can is empty beam 22022, trans runs 22024

[Gesner Passos]

Tester:

Test1:

• Open SANS INTERFACE (Interfaces->SANS->SANS ISIS)
• Select SANS2D instrument
• Load Mask (​http://trac.mantidproject.org/mantid/attachment/ticket/8245/MASKSANS2D_133R_12m_8mm_XPress_changer.txt)
• Check the 'Reduction Settings' tab you must find the TransFit line with selector set to BOTH and everything unselected.
• Change to batch-mode
• Load ​http://trac.mantidproject.org/mantid/attachment/ticket/8245/test_trans_inp.csv
• Run 1D reduction.
• Check History of workspace 22031rear_1D_1.5_12.5_can_tmp, the forth algorithm is RebinToWorkspace and the WorkspaceToRebin is 22027_trans_can_1.5_12.5_unfitted
• Check History of workspace 22031rear_1D_1.5_12.5_sam_tmp, the forth algorithm is RebinToWorkspace and the WorkspaceToRebin is 22027_trans_sample_1.5_12.5_unfitted

Test2:

• Open the file downloaded ​http://trac.mantidproject.org/mantid/attachment/ticket/8245/MASKSANS2D_133R_12m_8mm_XPress_changer.txt
• Uncomment the line: FIT/TRANS/LOG 1.5 12.5
• Check that 'Reduction Settings' is still set to both but, now to use the trans Fit and apply the logarithm method.
• Perform the Reduction again (1D Reduce)
• Check History of workspace 22031rear_1D_1.5_12.5_can_tmp, the forth algorithm is RebinToWorkspace and the WorkspaceToRebin is 22027_trans_can_1.5_12.5
• Check History of workspace 22031rear_1D_1.5_12.5_sam_tmp, the forth algorithm is RebinToWorkspace and the WorkspaceToRebin is 22027_trans_sample_1.5_12.5

Test3:

• Open the file downloaded ​http://trac.mantidproject.org/mantid/attachment/ticket/8245/MASKSANS2D_133R_12m_8mm_XPress_changer.txt
• Change the line: !FIT/TRANS/LOG 1.5 12.5 per FIT/TRANS/CAN/LINEAR 1.5 12.5
• Check that 'Reduction Settings' is set to Separate and the Can is configured to use Linear while the first line is set to not fit.
• Perform the Reduction again (1D Reduce)
• Check History of workspace 22031rear_1D_1.5_12.5_can_tmp, the forth algorithm is RebinToWorkspace and the WorkspaceToRebin is 22027_trans_can_1.5_12.5
• Check History of workspace 22031rear_1D_1.5_12.5_sam_tmp, the forth algorithm is RebinToWorkspace and the WorkspaceToRebin is 22027_trans_sample_1.5_12.5_unfitted

Any variant of these tests should behave analogously.

[Gesner Passos]

re #8245: sans trans fit option

Correct the isis_reduction_steps to process properly the FIT trans option. Bugs found:

1 - Setting an option for 'both' (can and sample) did not override previous settings for them alone. 2 - Returning from fitMethod return many variants (LIN, LINEAR, ...). It is better to return just one variant of the same method (LINEAR, LOGARITHMIC, and PolynomialX) 3 - Add a method to check if the trans was set using the specific selectors (sample and can).

Changeset: d68a5b12ca32a8aa25812b82ce8c2a0c0f0bc588

[Gesner Passos]

re #8245: sans trans fit option SANSRunWindow

Correct the SANSRunWindow to process properly the FIT trans option.

1 - Separate the loading Trans settings and the selector changing option. 2 - Correct the option to identify fit on/off 3 - Correct the parser of the fitMethod for CAN

Changeset: 76829d8ea6d9340c7c2cdc3ff27c6b3d9951549d

[Roman Tolchenov]

The tests seem to work but when I close the interface mantid crashes.

[Gesner Passos]

This issue is not directly connected to this ticket. The same thing occurs dealing with the master branch. Hence, it will be addressed in a separated ticket: #8294

[Roman Tolchenov]

I don't see the crash I've seen before.

[Roman Tolchenov]

Merge remote-tracking branch 'origin/bugfix/8245_sans_trans_option'

Full changeset: ​7b92714dfe2668eb57611ee656be753be55c210b

[Stuart Campbell]

This ticket has been transferred to github issue ​9090