Monitor Workspaces Not Being Replaced When New "Added" Event Mode Data is Assigned

[Peter Parker]

Richard has highlighted the following issue for the SANS ISIS interface:

• Reduce SANS2D run 23626, which will give us "23626_sans_nxs" and "23626_sans_nxs_monitors" workspaces.

• Add event mode runs 23626 and 23627 using the interface, giving us a "23626-add" file.

• Reduce this added-run file.

• Unbeknownst to users, the reduction is replacing "23626_sans_nxs" correctly, but not "23626_sans_nxs_monitors", which means that we're using incorrect monitor data.

A workaround he's being using in the meantime is to manually delete 23626_sans_nxs_monitors before he hit's "1D Reduce". We need to do this as part of the reduction so that this is not necessary.

[Peter Parker]

[Published to wrong branch. Deleted.]

[Peter Parker]

There are a few problems here:

1. For some reason our workspace naming convention has never distinguished between having loaded 23626 and having loaded 23626-add. Loading either of these will produce a workspace called 23626_sans_nxs, even though they are fundamentally different and should be named as such. I'm going to go out on a limb and say a change in this convention will break the reduction in some unforeseen ways.

2. 23626 is event data. 23626-add is a product of the "Add Runs" tab and so is histogram data. When loading event data, we produce a "NNNN_sans_nxs" workspace and a "NNNN_sans_nxs_monitors" workspace. When loading histogram data we only produce a "NNNN_sans_nxs" workspace, which contains the monitors as well as the data. This was always going to be problematic. We need to do things one way and then stick to it.

3. Our way of reading the monitors for a given workspace is to simply see if "NNN_sans_nxs_monitors" exists.

These things combine to cause the problem described by this ticket when we load 23626 and then 23626-add.

Another unrelated problem:

4. "23626-add" is another example of a bad naming convention, since we have no way of knowing exactly which runs it consists of. Does it contain 23627? Probably. Does it contain 23628, 23629 or 23630? Who knows.

[Peter Parker]

Refs #9429 - When loading histogram data, delete any monitor ...

... workspaces that may have been loaded for similarly-named event mode data.

Changeset: 87ad0949b95f8bdd1bd01c9350863a4c01eef9ca

[Peter Parker]

I've opened #9560 and #9561 to address some of the other issues.

Testing (ISIS only):

• Open the SANS ISIS interface and select SANS2D or SANS2DTUBES as the instrument to use.

• Load a user file. MaskSANS2DReductionGUI.txt from the system tests repo will do.

• Until #9550 is tested, you will have to manually add the SANS2D archive folder for the latest cycle (on Windows this is
  isis\inst$\Instruments$\NDXSANS2D\Instrument\data\cycle_14_1) to your list of managed directories.

• Type "23626" into the Scattering Sample field and then click "Load Data". You should see 23626_sans_nxs and 23626_sans_nxs_monitors in the workspace list.

• Now click the Scattering Sample browse button and select the only file in
  olympic\babylon5\Public\PGP\Test Data\9429. Clicking "Load Data" should leave you with only 23626_sans_nxs in the workspace list.

This is enough to ensure the changes work. If you want more details, please see me.

[Samuel Jackson]

Merge remote-tracking branch 'origin/bugfix/9429_monitor_workspaces_not_being_replaced_when_new_added_event_mode_data_is_assigned'

Full changeset: ​6d5bea345f514cdecc32574fa1547f55acb1af8b

[Stuart Campbell]

This ticket has been transferred to github issue ​10272