Reflectometry GUI Testing¶

Introduction ¶

The ISIS Reflectometry interface is a graphical front end for the main ISIS reflectometry reduction algorithm ReflectometryISISLoadAndProcess and post-processing algorithm Stitch1DMany. The purpose of the interface is to speed up the analysis of the Reflectometry spectra by providing batch-mode execution of these algorithms on groups of runs.

Set up ¶

These instructions assume use of MantidWorkbench.
Ensure you have the ISIS Sample Data in your Mantid user directories.
Ensure you have the ISIS archive enabled.
Open the ISIS Reflectometry interface.
Go to the menu Tools->Options and ensure all of the warnings are ticked. On the Rounding tab, set the precision to 2 decimal places.

Loading batch files ¶

Go to the menu Batch->Load.
Browse to the sample data and select file INTER_NR_test.json.
The interface should be populated with some information, including some content in the runs table.
Click the Expand All Groups button on the toolbar to expand the groups in the table.
The decimal values should be shown to 2 decimal places as per the Options that we set.
Re-open the same batch file. You should not be prompted with any warnings (unless you changed anything).

Adding and deleting batch tabs ¶

The interface should open with two batch tabs created by default.
Click the cross on any tab to delete it.
Check that if all batch tabs are deleted from the interface then a new one is automatically added.
Check that you can create a new batch tab using the menu Batch->New option.

Changing the Default instrument ¶

At the top-left and bottom-right of the interface is a drop-down menu called Instrument - it can switch between INTER, SURF, CRISP, POLREF, and OFFSPEC. Try changing it.
This should change the instrument for the whole of Mantid - check it by going to File->Settings in the MantidWorkbench main menu.
Reload the batch file. This should change the instrument back to INTER.

Processing Rows ¶

Re-load the batch file to reset the state.
Select a single row within a group with the mouse.
Click Process. Some output workspaces should be produced and the row should turn green. (You need to click off the row to see the green highlighting.)
The process bar along the bottom should go from empty to full.
Repeat this test for multiple rows (selecting using Shift and Ctrl while clicking with the mouse - a bit like in Excel)
Once you have processed all rows inside a group, the group header will turn pale green (paler than the rows to indicate the contents are complete but post-processing has not been done).

Processing Groups ¶

Re-load the batch file to clear the state and expand the groups.
Click on a group header line and clicking Process. This will process all rows in the group as well as post-processing the group itself.
The group header will turn green (the same shade as the rows).
Now click off the table to deselect everything, then click Process.
You will get a warning about processing the whole table - click ok.
This will now process anything that was left unprocessed.

Generating Plots ¶

There are two plotting buttons on the toolbar - they look like little plots. The first one plots the individual rows (the reduced output) and the second one plots the post-processed (stitched) output for groups.

Assuming the whole table is processed:

Select a group header:
- Plot the reduced data (first plot button) - this plots two lines for the individual reduced rows.
- Plot the stitched data (second plot button) - this plots a single line for the stitched output.
Select a processed row:
- Plot the reduced data - this plots one line for that reduced row.
- Plot the stitched data - this does nothing for a row.
We can also over-plot outputs from different groups. Select the whole table with Ctrl-A:
- Plot the reduced data - this plots 4 lines (reduced output for all rows).
- Plot the stitched data - this plots 2 lines (stitched output for both groups).

Polarisation Corrections ¶

Polarisation corrections settings are under the Experiment Settings tab in the interface. These should only work with CRISP, POLREF or OFFSPEC.

In the Runs tab set the instrument to INTER.
On the Experiment Settings tab, the Polarisation Corrections combo box should be greyed out.
Go back to the Runs tab and set the instrument to OFFSPEC.
Back on the Experiment Settings tab, the Polarisation Corrections combo box should now be enabled and the Polarization Efficiencies combo box should be disabled.
Select Parameter File from the Polarisation Corrections combo box. The Polarization Efficiencies combo box should still be disabled.
Switch to Workspace from the Polarisation Corrections combo box. The Polarization Efficiencies combo box should become enabled and show a list of all loaded workspaces.
Switch to FilePath from the Polarisation Corrections combo box. Polarization Efficiencies should now appear as a line edit. It should appear red for invalid paths and white for valid paths on your system.
Switch back to the ParameterFile setting from the Polarisation Corrections combo box.
Back on the Runs tab, delete all rows in the table (this can be done by pressing Ctrl-A and then Delete).
Note that this will leave an empty row. In that row enter run number 44956 and angle 0.4.
Check you can process the row and it turns green.

Search by experiment ¶

Clear all rows in the Runs table and set the instrument to INTER.
In the Search box on the left, enter Investigation Id 1120015 and Cycle 11_3.
Click Search and the results list will fill with all the runs for that experiment.
Try selecting some of the results and clicking the Transfer button.
The run numbers should be filled into the main Runs table following these rules:
- Runs with the same title will appear in the same group.
- Runs in the same group with different angles will appear on separate rows.
- Runs in the same group with the same angle will be combined into the same row, and shown as a sum e.g. 13460+13462.
- Runs that are highlighted blue are invalid and will not be transferred. Hover over them to see a message explaining why.
- If the runs are already in the table, they will not be transferred again (i.e. no duplicates).

Copy, Cut and Paste ¶

Copy, Cut and Paste are available from the toolbar buttons, the right-click content menu, or the standard Ctrl- C, X, and V shortcuts.

Note that it is very picky about pasting onto the correct destination (i.e. group onto group, row onto row etc.) and gives a confusing error message about “depth and size” if you get it wrong - this just indicates that the operation is not possible.

Re-load the test batch file and then test the operations listed below.

These should work:

Select a row, copy it, and paste it onto a different row.
Select 2 rows from one group, copy them, and paste onto 2 rows in another group.
Select a group and copy it. Paste it onto another group.
Copy a group. Deselect everything in the table before you paste. It should be pasted as a new group at the end.

These will give an error:

Copy a row. Deselect everything and paste. This fails because we don’t know which group to paste into.
Copy a row and try pasting onto a group. This could in theory append the row into the group but is currently not implemented.
Copy a group and try pasting onto a row. This is not possible.

Discarding changes ¶

Re-load the test batch file.
Edit the data in the table e.g. change an angle.
Close the interface using the x button at the top.
You should be warned that unsaved changes will be lost. Click Cancel and nothing should be lost.
Try again and click OK. The interface should close and discard your changes.
Repeat the test but instead of closing the interface, now attempt to re-load the batch file.
You should be warned that this will discard your unsaved changes. Again, the OK and Cancel options should work as expected.

Saving a batch ¶

Load the test batch file.
Make some changes to the table or settings that you can easily remember.
Go Batch->Save. A Save dialog will appear. Select a file to save to and OK it.
Close and re-open the interface. Note that you should not be prompted about discarding unsaved changes.
Load your saved batch file and check that the items you changed are restored.

Save tab ¶

Close the Reflectometry GUI and re-open it to ensure any previous settings have been cleared.
In the Search box on the left, enter Investigation Id 1120015 and Cycle 11_3.
Click Search and the results list will fill with all the runs for that experiment.
Select one of the results that is not highlighted blue (i.e. run 11934) and click the Transfer button.
In the main Runs table, click to process the row.
Go to the Save tab and hit Refresh. The workspaces list will contain all of the workspaces in the ADS.
Select a workspace in the list that starts with IvsQ.
Type a valid path into the Save path textbox.
Type something in the prefix field you’d like to use to identify the file. The files are saved in the form [prefix][workspace][ext].
In the File Format section, select Custom format (*.dat). Check that option Additional columns (includes Q resolution) is ticked but disabled, as it is not applicable.
Untick Header and Q resolution and set the separator to Comma.
Click Save and open the file that should have been saved to the save directory you specified. It should contain 3 columns of numbers, separated by commas.
Tick Q resolution and re-save. It should now contain 4 columns of numbers.
Double-click on a workspace name in the left list, e.g. IvsQ_binned_11934. The right list should be populated with parameters but be disabled.
Tick Header and the parameters list should be enabled. Select a couple of them, e.g. nperiods and run_start, and re-save.
- The file should now contain some header text starting with MFT.
- Amongst other things this text should contain the logs you selected, e.g. nperiods : 1 and run_end : 2011-10-21T13:32:03.
Try changing the separator to spaces or tabs and check that the 3 or 4 columns of numbers are separated using that separator.
Change the dropdown to ORSO Ascii (*.ort). The Header checkbox, separators and parameter settings are not applicable so they should be greyed out. The Additional columns (includes Q resolution) checkbox should be enabled.
Select a single IvsQ_binned workspace from the left list, e.g. IvsQ_binned_11934, and click Save. Open the .ort file that should have been created in your specified save directory. You should get a header at the top starting with ORSO reflectivity data file. There should be 8 columns of numbers with headings Qz, R, sR, sQz, lambda, slambda, incident theta and sincident theta.
Untick Additional columns (includes Q resolution) and re-save (the Q resolution checkbox should still be selected from the earlier steps). The file should now contain 4 columns of numbers with headings Qz, R, sR and sQz.
Untick Q resolution and re-save. The file should now contain 3 columns of numbers with headings Qz, R, sR.
Change the dropdown to ORSO Nexus (*.orb). The availability of the settings should be the same as they were for the ORSO Ascii (*.ort) format. Click Save and check that a file with extension .orb is saved out. This should be a Nexus file type, so will not be possible to view in a text editor. Use an HDF5 viewer to check that the contents of the file appear sensible.
Change the dropdown to 3 column (*.dat). All the settings should be greyed out as they are not applicable. Click Save to create the .dat file. You should get 3 columns of numbers separated by tabs (including a leading tab). At the top there is an integer indicating the number of lines in the data.
Change the dropdown to ANSTO, MotoFit, 4 Column (*.txt). The settings remain greyed out. Click Save to create the .txt file. You should get 4 columns of numbers separated by tabs (with no leading tab).
Change the dropdown to ILL Cosmos (*.mft). The settings remain greyed out apart from the parameters which should now be enabled. Click Save to create the .mft file. You should get 3 columns of numbers padded by spaces (including leading spaces). There should be a header starting MFT which includes any parameters you selected, the same as the Custom format.
In the Automatic Save section, tick the Save automatically on completion checkbox. This should enable the Include individual row outputs for groups checkbox. Try changing the selected file format - the Save multiple datasets to a single file checkbox should be disabled for all file formats apart from ORSO Ascii (*.ort) and ORSO Nexus (*.orb).
With either the ORSO Ascii (*.ort) or ORSO Nexus (*.orb) format selected, untick the Save automatically on completion checkbox. This should disable the Save multiple datasets to a single file checkbox, as it is only applicable when auto-save is selected for one of the ORSO formats.
Try entering a non-existent or invalid save path and then try to Save. You should get an error saying that the path is invalid.

Preview tab ¶

Go to the Reduction Preview tab.
Type INTER45455 into the Run input. Set the Angle to 1 and click Load. The instrument view plot should display the data on a detector with four banks. Note, with this dataset, we expect an error “Detector with ID…” to be thrown at this stage.
Go to the drop-down underneath the color scale next to the second (slice viewer) plot and select SymmetricLog10. This should allow you to see the counts on the slice viewer plot more clearly. You should see what appear as roughly four horizontal lines of data on the plot.
Going back to the instrument view plot, click the rectangle-select button above it and draw a single region that selects all detector banks. The selected detector segments should be summed and the result plotted on the slice viewer, appearing as a single line of data.
Reduce the size of your original region on the instrument view and check that multiple regions can be added to the plot. Check that when moving and resizing regions, the slice viewer plot is updated.
Check that you can delete regions from the instrument view by selecting them and pressing the delete key on your keyboard.
Make sure you have at least one region selected on the instrument view.
Click the rectangle select button above the slice viewer plot and draw a Signal region on the plot. A reduction will now be triggered for the selected spectra and the result plotted on the 1D plot.
Click the drop-down on the rectangle select button and select Transmission. Draw a transmission region onto the slice viewer plot. Then, in the same way, add one or more Background regions. The reduction should be re-run each time a region is added:
- You should see the tab quickly disable and re-enable.
- Another run of ReflectometryISISLoadAndProcess will be logged in the Messages bar.
- The 1D plot should update (although this is usually only noticeable when changes are made to the Signal region).
Check that moving and resizing regions triggers a re-run of the reduction.
Check that you can delete one of the Background regions by selecting it and pressing the Delete key on your keyboard.
Click the Apply button at the bottom right of the tab. The selected regions of interest should be populated in the lookup table on the Experiment Settings tab.
Back on the Reduction Preview tab, click the export button above the top left of the 1D plot. This should export a workspace called preview_reduced_ws to the ADS.
Right-click the workspace and select Show History:
- In the Algorithms list, expand ReflectometryISISLoadAndProcess.
- Click on ReflectometryReductionOneAuto and check in the right hand pane that the inputs for ProcessingInstructions, BackgroundProcessingInstructions and TransmissionProcessingInstructions correspond to the ranges of spectra you selected.
- Expand ReflectometryReductionOneAuto.
- Click on ReflectometryISISSumBanks and check that the input for ROIDetectorIDs matches the range of detector IDs you selected.
Back in the Reflectometry interface, go to the Runs tab. In the Process Runs table on the right-hand panel of the tab, enter Run INTER45455 and Angle 1 into the first child row. Click Process.
Compare plots of the preview_reduced_ws (from the Preview reduction) with IvsQ_binned_45455 (from the batch reduction). They should be the same.
The plots on the Reduction Preview tab are located within three dockable widgets. Check that the widgets can be undocked, re-docked, re-sized etc. without error or loss of functionality.