Keep OverlapDetector generic but normalize the getters for BedRecord and Interval

[nh13]

Alternate to #34

[codecov]

Codecov Report

Attention: Patch coverage is 95.40230% with 4 lines in your changes missing coverage. Please review.

Project coverage is 92.66%. Comparing base (2d71949) to head (cf99c00).

Files	Patch %	Lines
pybedlite/bed_record.py	66.66%	2 Missing ⚠️
pybedlite/tests/test_overlap_detector.py	96.55%	2 Missing ⚠️

Additional details and impacted files

@@                       Coverage Diff                       @@
##           cv_generic_overlap_detector      #36      +/-   ##
===============================================================
+ Coverage                        92.61%   92.66%   +0.05%     
===============================================================
  Files                                8        8              
  Lines                              609      641      +32     
  Branches                           110      116       +6     
===============================================================
+ Hits                               564      594      +30     
- Misses                              27       30       +3     
+ Partials                            18       17       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[clintval]

I like it but I would only ask we add property names to make the interface between BedRecord and Interval consistent and not add the zero_based_start and zero_based_open_end at this time.

[msto]

Why is GenericGenomicSpan necessary? StrandedGenomicSpan subclasses GenomicSpan, why not just use GenomicSpan here?

[clintval]

I kind of like it because it is explicit for the reader and shouldn't have any performance impact. Otherwise it might be overlooked that there is a tiny part of the OverlapDetector that does change behavior based on strand:

pybedlite/pybedlite/overlap_detector.py

Lines 305 to 307 in 49ff82c

	@staticmethod
	def _is_negative(interval: GenomicSpan) -> bool:
	return getattr(interval, "is_negative", False)

pybedlite/pybedlite/overlap_detector.py

Line 300 in 49ff82c

self._is_negative(intv),

[clintval]

What do you think about adding a negative property to BedRecord so we have fewer duplicated getters across both classes instead of introducing is_negative to both?

[clintval]

I kind of like it because it is explicit for the reader and shouldn't have any performance impact. Otherwise it might be overlooked that there is a tiny part of the OverlapDetector that does change behavior based on strand:

pybedlite/pybedlite/overlap_detector.py

Lines 305 to 307 in 49ff82c

	@staticmethod
	def _is_negative(interval: GenomicSpan) -> bool:
	return getattr(interval, "is_negative", False)

pybedlite/pybedlite/overlap_detector.py

Line 300 in 49ff82c

self._is_negative(intv),