re #15518

Process workspaces after a combined sequential fit over runs in which each
run involves a simultaneous fit over groups/periods. Results table should work
properly.

re #15518

If each run used a simultaneous fit across groups and periods, use the
data presenter to transform the results so that the results table is in the
expected format.

re #15518

In sequential fit, for each run, set up a simultaneous fit of all groups
and periods together.

To do: handle output

re #15518

re #15518

Won't build with this at the moment, it will later be replaced

re #15518

Lazy update them before handling any data change

re #15518

Need to have grouping and plot type as member variables of data presenter.
They should be updated on change in the interface.

re #15518

re #15518

It's part of the MuonAnalysis interface so should go with it, and is
not a reusable widget. Also it needs to go here so that it can use the
fit data presenter's shared loading/processing code rather than
duplicating it.

re #15518

createLoadAlgorithm doesn't create a load algorithm.
It actually creates a MuonProcess algorithm.
The loading in the sequential fit dialog is always done by LoadMuonNexus.

re #15518

The bug was observed when the lineedit lost focus, starting the search,
so then starting the search again went straight to the inspection stage
(as the text had not been modified since the previous search).

re #9963

The sequential fit dialog runs the file search in case the user
has edited the run input box. We need to wait for the search to
complete before getting the filenames.

Previously this was done in a busy while loop using
QApplication::processEvents(). This has been changed so that the
fit is only started once the search reports that it is done.

(We use a new signal fileInspectionFinished() because the existing
fileFindingFinished() is emitted "too soon", i.e. when the thread finishes
but while its member variables are still being accessed).

re #9963

re #9963

re #15223

Default behaviour was to use Mantid's default instrument, which is wrong
when this differs to the instrument set via the drop-down list on the
front tab of the interface. Changed this to use the same instrument as that
of the workspace in the fit browser.

re #15223

Default behaviour was to use Mantid's default instrument, which is wrong
when this differs to the instrument set via the drop-down list on the
front tab of the interface. Changed this to use the same instrument as that
of the workspace in the fit browser.

- If "ApplyDeadTimeCorrection" is set but there is no dead time table set
on the algorithm passed in, we need to load the dead times from the file.
- The output workspace type has been corrected.
- In the catch, write the exception's error message to the log.

The name of "MuonLoad" is now very confusing as it no longer loads data
from a NeXus file. Change it to "MuonProcess" as this better reflects what
it does.

This is also a good thing because the algorithm has changed quite a bit,
and so changing its name will avoid confusion with the old one.

Must now load the file separately using LoadMuonNexus, and use the loaded
workspace and time zero in MuonLoad.

(The name of MuonLoad needs to be changed because this is confusing now.
Perhaps MuonProcess would be better?)

Refs #9059

After the file searching thread is finished, it emits some signals
which MWRunFiles needs to process before we can access the found files.
We are now doing that by calling QApplication::processEvent() after the
search process has finished.

In that way it can be used safely in other places.

So that user is aware that something is happening.

Added an additional "preparation" state which is when file search is
done if necessary.