Spike qc nonburst - part 1

[ocehugo]

Hey guys,

this is the code for the Spike QC functionality that need to be merged for the Q3 release.

What's included:

• Some small cleanups
• Some refactoring/plumbing in previous routines to be more robust and permit burst processing in the future (as well as extra checks regarding time variables).
• The actual Spike Routines (Hampel,SavGol,etc).
• The SpikeQC driver with UI elements.

I poking @ggalibert since he will be in a better position to review the UI elements I think. It's pretty spartan atm, but I seek to improve it for the next spike processing (bursts). Most important thing to be done yet is to easy allow the users to change the filtering parameters. I tried my best to code some dynamical UI textboxes, but I didn't have enough time to finish and debug the callback hell :)

[ocehugo]

@ocehugo I think it would be quite valuable to add some unittest on the spikeQC function/wrappers to check the qc output on a given spiky array. It would help us validating the migration to a different version of matlab as well. What do you think?

Agreed. However, the tests can only be done at the functional level and not UI level. I got a half-baked suite here dangling with a mix of basic tests and realistic ones (from netcdfs) for the hampel method. I'll do some clean up, mix/match the tests among the available tests, and put it here. Some realistic tests I will leave for later because I don't want to delay this anymore and I also need to cherry-pick other variables.

The challenge here is to pick-up some good cases apart from artificial ones

[ocehugo]

comments addressed in the new commits @lbesnard

[ocehugo]

@lbesnard - I just add some extra commits fixing stuff that I encounter when trying to build the standalone packages and further issues when trying to mint the release.