Ticket #10598 (closed: fixed)

Opened 6 years ago

Last modified 5 years ago

SofQW algorithms discard the detector table

Reported by: Dan Nixon Owned by: Dan Nixon
Priority: critical Milestone: Release 3.3
Component: Indirect Inelastic Keywords:
Cc: spencer.howells@… Blocked By:
Blocking: Tester: Federico M Pouzols

Description (last modified by Dan Nixon) (diff)

The grouping information is lost and the first two spectra (assuming (OS)IRIS) are believed to be monitors, causing any further analysis relying on detector information to output garbage for the first two spectra and incorrect results for all others.

This happens for SofQW, SofQW2 and SofQW3.

Attachments

sqw_test.py (840 bytes) - added by Dan Nixon 6 years ago.

Change History

comment:1 Changed 6 years ago by Dan Nixon

  • Status changed from new to assigned

comment:2 Changed 6 years ago by Dan Nixon

  • Description modified (diff)

comment:3 Changed 6 years ago by Dan Nixon

  • Status changed from assigned to inprogress

Add proper detector mapping to SofQW

Refs #10598

Changeset: 56b6abdba1c91f4efedfe60f9d8ca89e82e356a4

comment:4 Changed 6 years ago by Dan Nixon

Build the spectra-detector map correctly

Refs #10598

Changeset: e406cc50a6cc2938a2d4efb9fabe5481bc25b53e

comment:5 Changed 6 years ago by Dan Nixon

Add proper mapping to SofQW2 and 3

Refs #10598

Changeset: 18c2c3cb84e3c9625f439ba39e8a157b9cf14c8d

comment:6 Changed 6 years ago by Dan Nixon

To test:

  • Run the attached script (requires systemtests/Data to be in data search paths).
  • Compare the detector tables of irs53664_graphite002_sqw and new_sqw, notice the first has some spectra mapped to monitors (this data contains no monitor spectra).
  • Compare the plots of original_sqw_conversion and new_sqw_conversion, notice the original has two obviously incorrect spectra (the ones mapped to monitors), the new one has much more sensible data.
  • Run the same script with he other two SofQW algorithms (SofQW2 and SofQW3), AFAIK there is no existing data to compare with so run the second half of the script on master then on the feature branch to compare results.
Last edited 6 years ago by Dan Nixon (previous) (diff)

Changed 6 years ago by Dan Nixon

comment:7 Changed 6 years ago by Dan Nixon

  • Status changed from inprogress to verify
  • Resolution set to fixed

comment:8 Changed 6 years ago by Dan Nixon

  • Priority changed from major to critical

comment:9 Changed 6 years ago by Federico M Pouzols

  • Status changed from verify to verifying
  • Tester set to Federico M Pouzols

comment:10 Changed 6 years ago by Federico M Pouzols

  • Status changed from verifying to closed

This also seems to work for me. Everything as expected:

  • irs53664_graphite002_sqw has the first two detectors marked as monitors. That doesn't happen in new_sqw.
  • The two first spectra of original_sqw_conversion are anomalous (kind of pulse). In new_sqw_conversion all the spectra look similar.
  • SofQW2 and SofQW3 produce similar results.

comment:11 Changed 6 years ago by Federico Montesino Pouzols

Merge remote-tracking branch 'origin/bugfix/10598_sqw_algos_discard_detector_table'

Full changeset: 1088d8588b4923db18655d0bf44bf7c6912406f1

comment:12 Changed 5 years ago by Stuart Campbell

This ticket has been transferred to github issue 11440

Note: See TracTickets for help on using tickets.