Add display average profile only as an optional display.

[githubdoe]

This was a request years ago. I added the average profile as a blue line to the show 16 diameters. However it is hard to see amongst the 16 diameter shown in smaller black lines. I have found it useful to see just the average. I have added a right menu option to the profile that shows only the average.

Previously we did not have space for that option along the top of the profile options. But now with the right click menu it is possible.

Currently it gives the option no matter what profile feature it is displaying. That is you can set it even in show single or show multiple wave fronts. But it only shows the average if you have selected show multiple diameters. That probably is confusing to users other than me.

I have the code in my local repro. I have yet to make a branch of it however. I will do that soon.

In addition I have asked AI to create a true average profile from all the points and not just from the 16 profiles like I did. The results is similar and I can't decide if I should add the new one. It presently only creates a profile from 0 to the +side. That code is commented out at present.

[githubdoe]

Also now that there is the right click menu we might be able to create a better interface for what is displayed based on what is selected in the wave front list.

Changing the options to display the current profile for what ever is selected, Display 16 diameters for what ever is selected, or display average profile for what ever is selected. The top button selectors would then go away.

[gr5]

It would be good to draw the average last such that is on top of the 16 diameters and thus easier to see.

[githubdoe]

I think it is drawn last. One of the things you probably don't know or have is that I enabled it to be drawn for each selected wave front. I use it to compare the average from the before and current session. That helps show the average difference between the two.

Here is an example with all but the average shown on the lower one. To bad I have not taken the time to rewrite the legend code for the all diameter case when now it can display more than one wave front.

[githubdoe]

I discovered always showing what is selected and deleting the show one and show multiple selected buttons simplifies the code and the user interface. I think that is going to be very nice.

[githubdoe]

I think my code rework of the profile displays and menu options is all working. There may be things I missed. I also changed the autoInvert dialog popup to default to conic. That at least for my own version. While we debate for final implementation. I hope some can give my additions a try. They are in branch averageProfile.

[githubdoe]

A note to self. It crashes when playing with the wave front list. Always when pressing select none. once when deleting a group in the middle.

A just analyzed igram creates a profile with a straight line from the right side up to 0. Pressing things that update it create the correct wave front. It is possilble now to have no profile displayed. Don't remember the conditions that happens.

[githubdoe]

Would really like a way to yoffset the whole 16 diameter group instead of dragging each one individually.

[gr5]

How about a gui where you right click a spot in the profile graph and select "bring profiles to this point" and it takes all the profiles and modifies the yoffset such that they all pass through that point on the graph.

If you do this on the left edge then they all match on the left edge. If you do this in the center they all match at the center. And so on.

The right click x,y point should be available somehow. Or if that's too hard to find you can have a context menu: "on next click in profile graph, bring all profiles to that point" or something shorter: "next click aligns profiles".

[gr5]

Or yet another checkbox. When checked then all clicks in the profile area move all the profiles vertically so they intersect with the mouse click.

[githubdoe]

The problem is to offset only one of the wave front groups and do the other at a different one. The GUI is not the problem there are several ways for it. But the offset at present happens in the QWTPlot logic and not in the profile plot code itself. The QWTPlot has each diagonal as a separate curve. So changing the code to find each curve group is a bit complex. More complex than I have time for at the moment.

The QWTPlot part is the code that watches the mouse press and movement.

[githubdoe]

It crashes when I'm selecting many wave fronts. In this case I was going to save them. Perhaps selecting many wave fronts tries to plot them all and crashes. Do I have a log file that I or others can look at that might help debug the issue?

[gr5]

I assume you are using qt creator and building in debug mode. you need to add a breakpoint.

main.cpp roughly line 37 in the myUnhandledExceptionFilter() function. I did that probably a year ago and never have to touch it.

If there is a crash then it will halt there and you can look at the stack trace in qt creator.

The log file is useless if you just built DFTF because you need a .debug file to interpret it

But yes there is a log file. It should also go into the "application output" window: at the bottom of qt creator there is "issues", "compile output" and others. Check "application output". The log file -- it will be in a subfolder of wherever your exe is located called DFTFringeLogs/

The log file is a pain to figure out. You need a ".debug" file. But sometimes there is an error from opencv to give you a hint of the error - something like "matrix accessed outside of bounds". But that's not much help.

[githubdoe]

No changes to code but it has not crashed in several days. I added a way to offset the 16 diameters by moving either the average or the original one profile. They stayed offset only until some new selection is applied to the wave front list. That is on purpose.
I have pushed all my current changes and the code branch could be tried out any time.

[githubdoe]

Added context menu to Ronchi image to create a ronchi image for each selected wave front. Currently I'm working on a mirror with Mel Bartels help as mentor. He likes to use Ronchi as his guide. I want to show how the ronchi has changed from session to session. So this feature does that for me. I had Gemini AI write it for me with a lot of suggestions on my part in trial and error. It took about 2 hours to get it to this current state. In the process we improved the code for Ronchi and Foucault so that it can be called from many different sources. That should improve displaying them in the reports.

[atsju]

Thank you Dale. Let us know whenever the code is ready for review and merge by openning a pull request.
It's easier to review and debug with several short pull requets than one large. Your changes are already consequent.