Use pyranges for overlap logic

[eharr]

Some of the overlap logic implemented in GenomeIntervalDf could be replaced with calls to the better documented and tested pyranges api.

[endrebak]

Actually, the NCLS is tested extensively in the pyranges tests, but indirectly.

[eharr]

Actually, the NCLS is tested extensively in the pyranges tests, but indirectly.
Ah sorry, I meant better tested than the pore-c implementation.

[endrebak]

Also, I would love to help out if you have reasonable feature-requests for pyranges. Or feel free to ping me for help :)