Modify existing output save formats to include Polarised SANS reduction outputs

[rbauststfc]

The existing NXcanSAS reduced data format needs to be modified to include relevant metadata from the Polarised SANS reduction at ISIS. The NXcanSAS standard is agreed by an international working group that determines the file format specifications, so our scientists will need to put together some proposed changes and have them agreed by the working group.

This issue is included in the MVP for the epic.

File Output Spec

See proposal for extending the NXcanSAS standard to support polarised data:

https://wiki.cansas.org/index.php?title=NXcanSAS_v1.1

We are still waiting for confirmation that the proposals have been accepted by the NXcanSAS working group.

Data Workflow

• The saving only needs to work for 2D data, and so we only need to look at saving and loading the NXcanSAS format.
• The data blocks need to be readable in any analysis software, but as long as we stick to the standard agreed by the working group, we should be fine on that front. We will want to keep SASView informed as we work on this, though, to ensure they can successfully load what we implement.

Requirements

The attached document captures information from requirements discussions with our scientists:

Polarised SANS NXcanSAS Metadata.docx

Agreed Solution Design

The below document details the final solution that we have discussed and agreed with our scientists:

Polarized SANS NXcanSAS metadata design.docx

Loading Polarised NXcanSAS Data

In addition to allowing data to be saved out, we will also need to provide the functionality for it to be loaded back into Mantid by amending the LoadNXcanSAS algorithm. I've only tested one file, but the current reloading of NXcanSAS files looks like it loads the data but not much metadata from the file. The instrument is not loaded onto the workspace, for example, and the sample logs are minimal. So a sensible first step may be to make sure we can load multiple data blocks rather than worrying about loading the additional metadata into the workspace. We can discuss with our scientists closer to the time.

[cailafinn]

From Steve K:

My tuppence would be to put the multiple periods in single output files unless there is a reason not to…

We know SasView will read multi-block files. However, I’m not sure that LoadCanSAS1D or LoadNXCanSAS do if you wanted to get them back into Mantid. So that loop would need closing.

My understanding is that they don't accept multiple blocks at present. So would need to modify the Savers and Loaders to save out the GroupWorkspaces properly into multiple blocks in the same file and vice versa.