Initial commit remove jsts for turf/difference

[rowanwins]

• Use a meaningful title for the pull request. Include the name of the package modified.
• Have read How To Contribute.
• Run npm test at the sub modules where changes have occurred.
• Run npm run lint to ensure code style at the turf module level.

Well I thought I'd make the much anticipated start of removing jsts from turf using martinez.

The initial commit covers turf/difference only.

So far

• Tests look good except for create-hole, this is a known issue in Martinez at the moment,I'll try and dig into this some more.
  • There is some minor coordinate shuffling that happens
  • I've left the jsts results in the package for the timebeing in a subfolder just while we play around
  • Note that Martinez also had issues with test case for issue Getting error while finding difference between a multipolygon and a polygon #721 if I removed the check on lines 47/48

Benchmarks

Generally martinez is smashing with jsts although there are a few exceptions around issue 721 for some reason...

 * ==using jsts==
 * clip-polygons x 2,512 ops/sec ±30.29% (71 runs sampled)
 * completely-overlapped x 6,777 ops/sec ±4.08% (78 runs sampled)
 * create-hole x 5,451 ops/sec ±9.92% (75 runs sampled)
 * issue-#721-inverse x 434,372 ops/sec ±3.40% (85 runs sampled)
 * issue-#721 x 421,194 ops/sec ±4.35% (80 runs sampled)
 * multi-polygon-input x 1,904 ops/sec ±5.16% (79 runs sampled)
 * multi-polygon-target x 1,240 ops/sec ±5.44% (79 runs sampled)
 * split-polygon x 2,468 ops/sec ±4.75% (82 runs sampled)

 * ==using martinez==
 * clip-polygons x 38,724 ops/sec ±11.98% (65 runs sampled)
 * completely-overlapped x 70,644 ops/sec ±8.29% (67 runs sampled)
 * create-hole x 62,128 ops/sec ±7.45% (75 runs sampled)
 * issue-#721-inverse x 354,473 ops/sec ±2.46% (81 runs sampled)
 * issue-#721 x 339,053 ops/sec ±3.24% (78 runs sampled)
 * multi-polygon-input x 17,728 ops/sec ±10.01% (64 runs sampled)
 * multi-polygon-target x 14,077 ops/sec ±6.52% (75 runs sampled)
 * split-polygon x 29,258 ops/sec ±8.99% (69 runs sampled)

Anyway off to pretty good start I think thanks to the incredible work by @w8r

[DenisCarriere]

🎉 Woot! Those benchmarks look amazing! Great job @rowanwins implemented this and thanks @w8r for the work on Martinez

[DenisCarriere]

😢 Same issue is occuring with Martinez?...

[rowanwins]

Yes the issue still occurs in Martinez but I didn't dig into it too deeply at 11pm so I'll look into it further

[DenisCarriere]

@rowanwins 🤔 ... you sure this is correct?... That was the initial issue was trying to handle the holes.

[DenisCarriere]

Seems like the Inner blue should be a lighter shade.

[rowanwins]

Correct - so martinez currently has an issue around this, I had a thought on how to resolve it last night lying in bed so I'll try and tackle it today

[rowanwins]

FYI I have a pull request in to resolve this issue upstream in martinez

[DenisCarriere]

@rowanwins Going to put this PR on the back burner until your PR w8r/martinez#55 gets resolved/merged.

[w8r]

Just merged w8r/martinez#55 and released on npm

[DenisCarriere]

👍 Great Thanks @w8r I'll update this PR with your new changes to Martinez.

[mfogel]

Edit: This appears to be a problem internal to martinez, part of w8r/martinez#35

I was checking out to see if this branch fixes #1224 and it appears that right now, this branch still has problems with the situation described by #1224. Here's a test case that can be dropped into turf-difference/test/in that causes node to crash with a heap-out-of-memory error (which I think is probably runaway recursion but I haven't verified that).

{
  "type": "FeatureCollection",
  "features": [
    {
      "type": "Feature",
      "properties": {},
      "geometry": {
        "type": "Polygon",
        "coordinates": [[[-1, -10], [0, -10], [0, 10], [-1, 10], [-1, -10]]]
      }
    },
    {
      "type": "Feature",
      "properties": {},
      "geometry": {
        "type": "MultiPolygon",
        "coordinates": [
          [[[-1, -10], [0, -10], [0, 10], [-1, 10], [-1, -10]]],
          [[[0, -10], [1, -10], [1, 10], [0, 10], [0, -10]]]
        ]
      }
    }
  ]
}

[rowanwins]

Thanks for the research and test case @mfogel - we'll take a look and see what we can do

[DenisCarriere]

👍 Looks like it's working now.

Going to publish this under v6.0

CC: @rowanwins @w8r

[DenisCarriere]

👍