ENH: enhance reliability of receiving EDF files at the end of an experiment

[mh105]

This PR contains

• a BF to unregister ioHub keyboard monitor of escape and space releases as discussed here.
• ENH commits to make receiving EDF files at the end of an experiment more reliable. Specifically, dependent on the merge of this PR into the main psychopy release, the current PR will use the exposed PsychoPy filename when dsfile is None to open and download EDF files with the same naming convention as the rest of subject data files. If the saving folder does not exist, we also mkdir to make sure we do save the eye tracking data.
• BF around dependency of SR research package sr-research-pylink from their own PyPI server.

[mh105]

Apparently the PsychoPy plugin environment manager is quite general as shown here. But this generic formulation is notably a problem in this situation with a non-PyPI index for installing a dependency of the plugin. Perhaps you just have to use an if statement to handle this edge case! I'll leave that to the development team to decide how to handle...

[TEParsons]

I think the solution is going to be us pestering the eyelink team to put their package on PyPi :P

I've pulled this in - @mh105 would you mind checking with the latest psychopy/dev and main branch of this repo installed whether your eyelink is now found and calibrates correctly? If so, I'll queue up a release of this plugin to coincide with the next release of PsychoPy and all these fixes can be live :D

[mh105]

In order for some of my commits to work correctly, I need the merge of this PR into the PsychoPy release and then into dev: psychopy/psychopy#6555

[mh105]

@TEParsons Just a gentle reminder to see if you could merge my other PR into PsychoPy release and dev branches so I can test out whether my EyeLink can be imported correctly using the new importlib approach. Thanks!

[mh105]

I've tested them - didn't work initially but after psychopy/psychopy#6610 and #10, they work smoothly now 👍