[Osiris] Indirect Diffraction Support

[Peter Parker]

It has been reported that the Indirect Diffraction interface is no longer working with Osiris data, where previously it had been seen working last June 2011.

Further comments from user:

[The] gui for iris and tosca are different from the Osiris one. The Osiris one will merge together several different d spacing runs, divide them through van runs and not through the monitor and use a cal file, which is different from the vanadium calibration file. This diffraction reduction is based on an approach used by the professional diffraction people.

The Osirisdiffred routine is asking for these input files, which is correct. However it does not merge file together, this was then implemented into the gui.

[Peter Parker]

Refs #4820 - Fix for OSIRISDiffractionReduction algo.

Fixed a bug where we were expecting 5 files exactly. Have forwarded the algorithm to the user for testing; if he is happy with it I will then rig up the interface to use it if Osiris is selected in the drop down box.

Changeset: 3dccec9e7f79fbcccf946a7af129083510d02bb0

[Peter Parker]

A more complete set of requirements from user:

• Set up the GUI under the assumption that the number of vanadium and sample files are identical.
• Procedure for vanadium and sample:
  • Load
  • Normalise to current
  • Align detectors
  • Diffraction focusing
• Divide through vanadium
• Identify dranges with tof regime
• Merge files
• Output: in drange and tof signal (so conversion back into tof)

Next step is to allow more flexibility in input files for the sample:

• Practically the number of sample files is often different from the number of vanadium files and there can be several from the same range, so averaging is necessary before the reduction continues.
• It might be worth at that point to consider adding to the Osiris parameter file the vanadium runs and the cal file, because these ones will not change during a whole cycle normally. Only the sample input is varied. These stored files could then automatically appear in the gui and are no more necessary to load.

(Note: It would be trivial to instead remember cal and van files via QSettings. Need feedback from user to see if would be an acceptable alternative.)

Later there might be 2 further steps necessary: masking detectors and absorption corrections. No priority for the moment.

[Peter Parker]

Refs #4820 - Workspace name suggester for multi files.

Added a "suggestWorkspaceName" function to the MultiFileNameParsing namespace, which takes in a vector of filenames, and returns a string which we can use as a workspace name for the loaded files - assuming they are to be added or merged together in some way to form a single workspace.

Changeset: bf13de59b84b0149e918df73abbe3190a2dd53aa

[Peter Parker]

Refs #4820 - Adding further OSIRIS drange values.

Allows a full range of runs to reduced by the OSIRISDiffractionReduction algorithm. Currently, the range must still start with the first run, and carry on uninterrupted up to the last run specified.

Changeset: 298245b30e0fd410ea71408b0085526596503078

[Peter Parker]

Refs #4820 - Further changes to InDiff interface.

Van files and cal files are "remembered" by GUI. Small modifications to the interface of the OSIRISDiffRed algo.

Changeset: 79257ed8b5b4c94cf83fbca95b6670708d113312

[Peter Parker]

Refs #4820 - Possible fix for Lucid and RHEL6 builds.

std::make_pair does not seem to have overloaded versions for const paramters in gcc.

Changeset: b9edbd54f85bba01efa31021e9c0fb341edab0d8

[Peter Parker]

Refs #4820 - Further fix, but for Fedora and Oneiric builds.

Removing template arguments from call to std::make_pair.

Changeset: 574218cc7358756ba79403921a54eb1f2c9f2224

[Peter Parker]

Refs #4820 - Verbose, but definite fix this time.

Changeset: 83323dffa6c14bc7ef23f653300c3479d7cb44a6

[Peter Parker]

Refs #4820 - Now changing other instance of make_pair.

Changeset: 4b0bcee01a382808cc75f6250417a90b069cda7a

[Peter Parker]

Refs #4820 - Fulfil all initial OSIRISDiffRed requirements.

Main reduction is carried out fully, but now with more flexible input options - as long as there are vanadium files covering the DRange of every sample file, files will be averaged out where needed and the reduction will carry on as it should do.

UI now fully rigged up to algorithm.

Changeset: fcdf412292c719874bc303e3ffd90d2128d48a1b

[Peter Parker]

Refs #4820 - Add import of Reducer back into file to fix build.

Raising a separate ticket (#4924) for Martyn to look into this problem.

Changeset: f45037f6122c03b88e2a6de5c88dcd653b520d6a

[Peter Parker]

Refs #4820 - Update system test and test result.

Changeset: 5ec0b74e13a8282a98dc42713dd18f6ffcc6bfa0

[Peter Parker]

Refs #4820 - Workspace name suggester for multi files.

Added a "suggestWorkspaceName" function to the MultiFileNameParsing namespace, which takes in a vector of filenames, and returns a string which we can use as a workspace name for the loaded files - assuming they are to be added or merged together in some way to form a single workspace.

Changeset: bf13de59b84b0149e918df73abbe3190a2dd53aa

[Peter Parker]

Refs #4820 - Adding further OSIRIS drange values.

Allows a full range of runs to reduced by the OSIRISDiffractionReduction algorithm. Currently, the range must still start with the first run, and carry on uninterrupted up to the last run specified.

Changeset: 298245b30e0fd410ea71408b0085526596503078

[Peter Parker]

Refs #4820 - Further changes to InDiff interface.

Van files and cal files are "remembered" by GUI. Small modifications to the interface of the OSIRISDiffRed algo.

Changeset: 79257ed8b5b4c94cf83fbca95b6670708d113312

[Peter Parker]

Refs #4820 - Possible fix for Lucid and RHEL6 builds.

std::make_pair does not seem to have overloaded versions for const paramters in gcc.

Changeset: b9edbd54f85bba01efa31021e9c0fb341edab0d8

[Peter Parker]

Refs #4820 - Further fix, but for Fedora and Oneiric builds.

Removing template arguments from call to std::make_pair.

Changeset: 574218cc7358756ba79403921a54eb1f2c9f2224

[Peter Parker]

Refs #4820 - Verbose, but definite fix this time.

Changeset: 83323dffa6c14bc7ef23f653300c3479d7cb44a6

[Peter Parker]

Refs #4820 - Now changing other instance of make_pair.

Changeset: 4b0bcee01a382808cc75f6250417a90b069cda7a

[Peter Parker]

Refs #4820 - Final clean up of make_pair fixes.

Incorrectly diagnosed the problem with the make_pair errors. The first instance of make_pair was okay to be left in, just minus the template parameters. The second instance was unnecessary anyway, as the constructor could be used to init the object in that case.

Changeset: 921f1aaa1d79efc6f23a8a68aeb416f6295c8fe1

[Peter Parker]

Refs #4820 - Fulfil all initial OSIRISDiffRed requirements.

Main reduction is carried out fully, but now with more flexible input options - as long as there are vanadium files covering the DRange of every sample file, files will be averaged out where needed and the reduction will carry on as it should do.

UI now fully rigged up to algorithm.

Changeset: fcdf412292c719874bc303e3ffd90d2128d48a1b

[Peter Parker]

Refs #4820 - Add import of Reducer back into file to fix build.

Raising a separate ticket for Martyn to look into this problem.

Changeset: f45037f6122c03b88e2a6de5c88dcd653b520d6a

[Stuart Campbell]

Could you point me at an appropriate data file, cal file and vanadium file ?

Cheers Stu

[Peter Parker]

Van file 1

[Peter Parker]

Van file 1

[Peter Parker]

Van file 3

[Peter Parker]

Cal file

[Peter Parker]

Attachment file size limit reached, so will send you the other three files via email.

Normal usage is as follows:

• For all runs in the same cycle, the same calibration file and the same set of vanadium files are used. Once browsed to, the interface should remember them from then on.

• Each vanadium file corresponds to a certain DRange, as do the data files. The interface should stop and warn the user if they have asked to reduce a data file without providing the corresponding vanadium file. Providing an unnecessary vanadium file should not be a problem, the reduction should just ignore it and carry on regardless.

• It should be possible to load two or more data files with the same DRange. In this instance, an "average" is produced.

Vanadium Files: osi92257.raw to osi92259.raw. These cover the 1st, 2nd and 3rd DRanges, though there are actually 11 in total.

Calibration File: osiris_041_RES10.cal

Data Files: osi89770.raw to osi89772. These also cover the 1st, 2nd and 3rd DRanges.

Final output should two be 2 workspaces and their plots, if selected. (Note, #4852 has been raised to fix the error messages.)

[Stuart Campbell]

Tested using both the Indirect Diffraction GUI and the OSIRISDiffractionReduction algorithm.

[Stuart Campbell]

This ticket has been transferred to github issue ​5667