Improve @turf/boolean-overlap performance on MultiPoints

[mfedderly]

Changes:

• Don't make a copy of the multipoint coordinates
• Exit early when we know there is an overlap, instead of continuing on
• Removed some unused imports as a bonus

I wanted to try something similar on the lines and polygons, but segmentEach() doesn't have a way to exit early so I skipped it for now.

Here's the relevant bench.js output, I removed the ones for lines and polygons:

Before:

equal-multipoints: 0.271ms
multipoints: 0.231ms
multipoints: 0.025ms
single-multipoints: 0.021ms
equal-multipoints x 65,977 ops/sec ±1.40% (87 runs sampled)
multipoints x 1,664,734 ops/sec ±2.02% (83 runs sampled)
multipoints x 1,572,537 ops/sec ±1.52% (86 runs sampled)
single-multipoints x 1,860,378 ops/sec ±1.86% (89 runs sampled)

After:

equal-multipoints: 0.186ms
multipoints: 0.035ms
multipoints: 0.017ms
single-multipoints: 0.014ms
equal-multipoints x 69,743 ops/sec ±1.49% (88 runs sampled)
multipoints x 4,899,999 ops/sec ±1.88% (79 runs sampled)
multipoints x 5,212,289 ops/sec ±1.44% (79 runs sampled)
single-multipoints x 6,440,465 ops/sec ±1.53% (87 runs sampled)

[rowanwins]

Makes sense!