Turf: Bug: turf/intersect 6.1.3 fails due to martinez.intersection returning invalid polygon

[jomel]

This is likely related to #1394

martinez.intersection may return a list of coordinates that is not a valid polygon.
see here for the polygons causing issue: https://gist.github.com/jomel/fd12af66b2d8519f0419015fac563e7d

const intersect = require('@turf/intersect').default;
intersect(polygon1, polygon2);

results in
Error : 'Each LinearRing of a Polygon must have 4 or more Positions.'

I've narrowed it down to
let intersection = martinez.intersection(geom1.coordinates, geom2.coordinates);
https://github.com/Turfjs/turf/blob/master/packages/turf-intersect/index.ts#L52
returning

[[[[[[0.656784,51.561146],[0.657001,51.560858],[0.657391,51.560977],[0.657562,51.560767],[0.657145,51.560639],[0.657238,51.560525],[0.6568784588933171,51.560518700155825],[0.656997802573544,51.560360135547825],[0.657207,51.560361],[0.657205,51.560501],[0.657899,51.560714],[0.657584,51.561104],[0.65779,51.561168],[0.658176,51.56069],[0.658606,51.560818],[0.658553,51.560951],[0.658552,51.560955],[0.658551,51.560958],[0.65855,51.560962],[0.65855,51.560976],[0.658551,51.560979],[0.658552,51.560983],[0.658553,51.560986],[0.658555,51.56099],[0.658557,51.560993],[0.658876,51.561546],[0.6584919499757166,51.56157188198154],[0.658501,51.561511],[0.658495,51.561501],[0.6584314599374823,51.5614094720528],[0.658483,51.561406],[0.658465,51.561311],[0.658258,51.560948],[0.658022,51.561241],[0.657832,51.561182],[0.657747,51.561287],[0.657399,51.561179],[0.657306,51.561299],[0.656784,51.561146]]]]]]

which seems to have one or two sets of [ ] too many? Yeah, I know it's a problem with martinez, but can't turf be smart enough to handle this?

In my case swapping the order of polygons in the call to intersect() does not solve the problem (as it id in #1394 )

[ilijaz]

I've got problems with martinez too. Look at https://github.com/Turfjs/turf/issues/1482.
@rowanwins said there

In v7 I'm looking at swapping out martinez with the polygon-clipping library (which started as a fork of martinez), it's a bit slower but seems a bit more robust for most things.I'll take a look at your test case and see how it performs...

So you could try...

--

[rowanwins]

Hi @jomel & @ilijaz

This is resolved in [email protected]

Cheers

[arthursp]

Looks like there is the same problem for the booleanContains function

[valaxy]

where can I find [email protected]?