Allow curve editor to treat multiple IV-curve files

[michele-riva]

This came originally from a request from @jbahlmann. Here is a summary of an email exchange between @jbahlmann, @fkraushofer, @michele-riva, and @schmid-iap.

@jbahlmann: Fri, Apr 30 2021 Regarding the calculation of beams with imageJ: My experimental data are divided into two seperate sets due to degradation and issues with an energy dependant incident angle. What do you recommend, how to treat those data? I could just put the images together, but then there is a shift in positions, beam current, intensities,... . I am not sure whether this is an good solution. Maybe evaluation of the single parts and merging the resulting EXPBEAMS files is a better approach. But then I would need to treat the overlapping energy range somehow to have a smooth transition. Do you have different ideas?

@fkraushofer Hmm, interesting question, probably something for @schmid-iap / the curve editor. Yes, I would say a smooth transition from one to the other is best.

@michele-riva: A smooth transition would be ideal. Since we're doing something like this already in the curve editor [ndr: for treating those symmetry-equivalent spots that may not have intensities over the whole energy range of other symmetry-equivalent beams], one could use the same merging algorithm. Probably @schmid-iap is the best person to address this in detail. I don't know if he has something in place (or planned) for handling multiple videos, but probably doing a smooth transition between curves that have already been extracted from the data and normalized to the beam current should not be as much complicated.

@schmid-iap Appending two stacks does not seem a good option - there will be a jump of intensities (obviously, R_Pendry is sensitive to this). I think the best option would be to have some overlap (say, 20 eV). Then I could think of creating a utility that smoothly merges the I(V) curves, maybe taking a factor (average ratio in the overlap region) between the two measurements into account. This is probably better than smoothly merging the image stacks: "Energy dependent incident angle" sounds like the spot positions won't be the same and smooth merging would result in elongated spots where the two image stacks would overlap.

[michele-riva]

This tool would also be a good addition for LEEM-IV measurements performed by @AlesSala. In fact,

@michele-riva He says that one needs to acquire IV movies in a couple of chunks (e.g., 20-100 eV + 90-450 eV) as the MCP voltage has to be adjusted in between. I suggested that one extracts IV curves from the two (or more) chunks with the tracker, and then post-process them in the curve editor. This is more or less what @jbahlmann already asked, so it is probably good if the curve editor would allow one to load multiple 'raw' files.

This is currently in the works:

@schmid-iap A utility for stitching I(V) curves with some overlap is planned. It would adjust the intensities of the 2^nd measurement such that the sum of all intensities where there is an overlap is the same for the first and 2^nd series of measurements. Then it should apply a smooth transition in the x-axis (energy) range describing the overlap.

[schmid-iap]

Release v0.5.3 adds Average I(V) Curves and Stitch I(V) Curves in the ImageJ Plugins>ViPErLEED menu. For stitching, if there is some overlap, a smooth transition is used. (The weight is a linear function from 0 to 1 in the overlap region.) Since some beams may have almost zero intensity in the overlap region, there is one common scaling factor for all beams: For each dataset, the sum of all (non-negative) intensities (restricted to data also present in the other data set) is determined for the overlap region. The ratio of these sums is used for normalizing the second data set.