Hello, I thought I would post this here so it's a wider discussion
Would it be possible/sensible to change how CD_PROJECTION etc are defined, based on the dimensionality of the incoming data?
This way we could use a scanned application definition for a single point on say, a fluorescence scan, but it wouldn't matter which order the axes were scanned for tomo. ie. scans of x,y,theta or x,theta,y would all be equally valid
We could define them according to the NXtransformations
I just wanted to avoid having to write more "types" if we can reuse the existing ones....
Ah, you have spotted exactly the reason for the current planning for a rewrite we are doing at the moment, to deal with exactly this problem. Thanks for the info though, and I will keep you updated on changes made to the system
I have a note... Aaron earlier suggested here as a place for leaving them...
Because of the way hdf5 works, all output files are created at the beginning of the run and will be populate with incorrect meta_data if any changes are made during the processing. Does that sound correct and is it something we need to change?
Following on from our discussion today, I think the use of metadata now is a good step forward in this.