[GEO] OGC compliant polygons fail with ambiguity

[nknize]

The GeoJSON specification (http://geojson.org/geojson-spec.html) does not mandate a specific order for polygon vertices thus leading to ambiguous polys around the dateline. To alleviate ambiguity the OGC requires vertex ordering for exterior rings according to the right-hand rule (ccw) with interior rings in the reverse order (cw) (http://www.opengeospatial.org/standards/sfa). While JTS expects all vertices in cw order (http://tsusiatsoftware.net/jts/jts-faq/jts-faq.html). Spatial4j circumvents the issue by not allowing polys to exceed 180 degs in width, thus choosing the smaller of the two polys. Since GEO core includes logic that determines orientation at runtime, the following OGC compliant poly will fail:

 {
  "geo_shape" : {
    "loc" : {
      "shape" : {
        "type" : "polygon",
        "coordinates" : [ [ 
          [ 176, 15 ], 
          [ -177, 10 ], 
          [ -177, -10 ], 
          [ 176, -15 ], 
          [ 172, 0 ],
          [ 176, 15] ], 
          [ [ -179, 5],
            [-179, -5],
            [176, -5],
            [176, 5],
            [-179, -5] ]
        ]
      }
    }
  }
}

One workaround is to manually transform coordinates in the supplied GeoJSON from -180:180 to a 0:360 coordinate system (e.g, -179 = 181). This, of course, fails to comply with OGC specs and requires clients roll their own transform.

The other (preferred) solution - and the purpose of this issue - is to correct the orientation logic such that GEO core supports OGC compliant polygons without the 180 degree restriction/workaround.

[nknize]

Follow feature branch feature/WKT_poly_vertex_order for WIP

[nknize]

Fixed in #8762