[Feat] add selection and brushing for Geojson layer

[lixun910]

This PR is to add selection and brushing for Geojson layer:

1. add geometric centroids /mean centers to geojson layer

• mean centers for polygon and multipolygon
• mean centers for linestring and multiLineString (interpolate point at 50% length)

2. create a spatial index using geometric centroids

• Flatbush index

3. enable polygon filter on geojson layer; use spatial index to prefilter the datacontainer

http://www.faqs.org/faqs/graphics/algorithms-faq/

Performance tests:

• data: ~1 million polygons Utah house footprint

• centroids/mean centers

• Spatial index

Flatbush: 163ms

Summary

C++ code is normally 2x - 6x faster than the JS code with the same implementation. Boost::Geometry C++ has the fastest implementation. However, the centroid function in GEOS is not optimized and is slow (see explanation here Mean center and centroid)

• 1 million polygons Utah house footprint

  • centroids/mean centers
    • c/c++:
      • GEOS (Mass Centroids): 1112 ms
      • Boost (Mass Centroids): 21 ms
      • Mean Centers: 61 ms
      • Mass Centroids: 398 ms
    • js:
      • polylabel: 8036 ms
      • Mean Centers: 233 ms
      • Turf Mean Centers: 446ms
      • Turf Mass Centroids: 778 ms

• 3085 US counties

  • c/c++
    • GEOS: 56 ms
    • BOOST: 0.20 ms
    • Mean Centers: 0.56 ms
    • Centroids: 25 ms
  • js
    • Mean Centers: 2.25 ms
    • Turf Mean Centers: 5.2 ms
    • Turf Mass Centroids: 56 ms

Note: Why the GEOS Centroid() function is slow?

The GEOS centroid() function performs an additional calculation for the fallback case when the polygon area is 0. This additional calculation is quite expensive, as it computes the total length of all the edges and returns the point at length/2 as the centroid.

This PR is to add polygon filter based on mean centers for GeoJsonLayer. Mean centers are easy to compute (much faster than computing the mass centers or geometry centers), even though mean center is more affected by the points that are far away from the center of the shape (see notes below).

Notes:

Mean center vs centroid (mass center):

Mean center and centroid are two different ways to represent the center of a geometric shape.
The mean center is the average of all the points in the shape. To compute it, you add up the x-coordinates of all the points and divide by the number of points, and then do the same for the y-coordinates.

The centroid (a.k.a. the center of mass, or center of gravity) of a polygon can be computed as the weighted sum of the centroids of a partition of the polygon into triangles. The centroid of a triangle is simply the average of its three vertices, i.e., it has coordinates (x1 + x2 + x3)/3 and (y1 + y2 + y3)/3. This suggests first triangulating the polygon, then forming a sum of the centroids of each triangle, weighted by the area of each triangle, the whole sum normalized by the total polygon area.

In general, the mean center and centroid will be different. The mean center is more affected by the points that are far away from the center of the shape, while the centroid is more affected by the points that have a large area.