Ticket #7548 (closed: fixed)

Opened 7 years ago

Last modified 5 years ago

Fix issues with Quantification

Reported by: Martyn Gigg Owned by: Martyn Gigg
Priority: critical Milestone: Release 3.0
Component: Direct Inelastic Keywords:
Cc: Blocked By:
Blocking: Tester: Gesner Passos

Description (last modified by Martyn Gigg) (diff)

Russell Ewings has compiled some scripts & plots detailing differences with Mantid & TobyFit. These need to be addressed. The scripts and plots are attached

Attachments

Tobyfit_Mantid_differences_summary.pdf​ (49.4 KB) - added by Martyn Gigg 7 years ago.
iliad_merlin_sr122.py​ (1.2 KB) - added by Martyn Gigg 7 years ago.
Sr122_individual_components.py​ (8.2 KB) - added by Martyn Gigg 7 years ago.
Sr122_individual_components_plots.py​ (9.6 KB) - added by Martyn Gigg 7 years ago.
Sr122_individual_components_mg.py​ (6.9 KB) - added by Martyn Gigg 7 years ago.

Change History

comment:1 Changed 7 years ago by Nick Draper

  • Component changed from Framework to Direct Inelastic

comment:2 Changed 7 years ago by Martyn Gigg

  • Description modified (diff)

Changed 7 years ago by Martyn Gigg

Changed 7 years ago by Martyn Gigg

Changed 7 years ago by Martyn Gigg

Changed 7 years ago by Martyn Gigg

comment:3 Changed 7 years ago by Martyn Gigg

  • Status changed from new to inprogress

Enable IkedaCarpenterModeratorTest

Refs #7548

Changeset: 3ef44faec80129892fd2fc602dc3424fdc16f369

comment:4 Changed 7 years ago by Martyn Gigg

Remove static qualifier as it's a bit over the top.

Refs #7548

Changeset: b71059db6be4ec30ea6e1effb36ded7c3fb4a76b

comment:5 Changed 7 years ago by Martyn Gigg

Report progress message at beginning of quantification simulation.

Avoids the confusion where nothing seems to happen. Refs #7548

Changeset: 1a23a013cce05065fc06dab0b8695f27009e8921

comment:6 Changed 7 years ago by Martyn Gigg

Fix memory leak with random number generators.

Refs #7548

Changeset: 92014eb30eedbb8d73f1a25311dbf3242cadce3b

comment:7 Changed 7 years ago by Martyn Gigg

  • Milestone changed from Release 3.0 to Release 3.1

comment:8 Changed 7 years ago by Martyn Gigg

Fix crash with default TobyFitResolutionModel setup

Refs #7548

Changeset: 116615f0390840b5bbca53cb3723742f9538a4da

comment:9 Changed 7 years ago by Russell Taylor

Re #7548. Fix OS X builds.

I don't know if this can overflow a size_t anywhere, but the progress reporting going awry isn't such a big deal anyway.

Changeset: 586b24896a30b025e94359ab502ba53b431ae2e0

comment:10 Changed 7 years ago by Martyn Gigg

Fix error in TobyFit B Matrix calculation of sample rotation.

The sample-only contribition now looks much more like TobyFit. Refs #7548

Changeset: 29132ac088e594ca7b165a47815b8e009241b9e5

comment:11 Changed 7 years ago by Martyn Gigg

Fix bug in TobyFit mosaic calculation.

Refs #7548

Changeset: ae99fb0e4017a99a415cd610c4c9514ec4639832

comment:12 Changed 7 years ago by Martyn Gigg

  • Milestone changed from Release 3.1 to Release 3.0

comment:13 Changed 7 years ago by Martyn Gigg

Update TobyFitSimulation results after fixing some bugs.

Refs #7548

Changeset: 8eefe5d76373c66e537b3081ec6a0d2e9d9d0979

comment:14 Changed 7 years ago by Martyn Gigg

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

Branch: bugfix/7548_quantification_fixes in BOTH systemtests & main code repository

Tester: The raw data for this is all in the ISIS archive in MERLIN cycle_09_1, add this to your search path. Three additional files are required that can be obtained from

Download these to a directory in your search path.

I have attached a cut down version of the original Sr122 script plus the additional mapping files etc. The iliad script needs to be run first to generate the w1 workspace then run the Sr122_individual_components_mg.py script.

Plot each of the final matrix workspaces: mt_mod_sim,mt_volume_sim,mt_mosaic_sim and then compare them with the tobyfit ones in the attached pdf. They should be much more similar in shape that they orignally were.

Last edited 7 years ago by Martyn Gigg (previous) (diff)

Changed 7 years ago by Martyn Gigg

comment:15 Changed 7 years ago by Martyn Gigg

  • Status changed from verify to reopened
  • Resolution fixed deleted

comment:16 Changed 7 years ago by Martyn Gigg

  • Status changed from reopened to inprogress

Fix off-by-one error in IkedaCarpenterModerator model.

If the value from the sample table is nearer the last edge then simply returns the last value. Refs #7548

Changeset: 9069db198a05ee4194046e1c8c17181d34dfbbaf

comment:17 Changed 7 years ago by Martyn Gigg

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

comment:18 Changed 7 years ago by Gesner Passos

  • Status changed from verify to verifying
  • Tester set to Gesner Passos

comment:19 Changed 7 years ago by Gesner Passos

  • Status changed from verifying to closed

Merge remote-tracking branch 'origin/bugfix/7548_quantification_fixes'

Full changeset: ​b6196828f80479de4e9103f2d6cf75ec9bd0fdc6

comment:20 Changed 7 years ago by Gesner Passos

Merge remote-tracking branch 'origin/bugfix/7548_quantification_fixes'

Full changeset: ​5e1582579b0cfef52a5d22aa369cb762b71f8e03

comment:21 Changed 7 years ago by Gesner Passos

The solution is better right now, but, it is impossible to check if it is correct (by code review). The solution is mainly in the commit pointed in comment:10. But, there is not way to check if the indexes are correct. Because, the only point in the documentation that provides some information on the origin of this algorithm is this one: """Defined in Toby G Perring's thesis pg 112, equation A.48.""", which I could not find. Is is worth bringing the equation to the documentation?

But, it definitely improved so, ticket pass.

comment:22 Changed 5 years ago by Stuart Campbell

This ticket has been transferred to github issue ​8393

Note: See TracTickets for help on using tickets.