Method Geometry#buffer not defined for spherical factory about rgeo HOT 5 CLOSED

Another curious observation is that this Mercator factory returns distance results that are quite different from what Google Maps does. Spherical (and PostGIS geog) are much closer. I'm relatively new to GIS work, so forgive me if this is obvious information.

from rgeo.

The spherical factory actually omits a number of operations. This is simply because many geometric operations are extremely difficult to compute on a curved domain. Even PostGIS omits some and approximates others. At this point, there aren't immediate plans to implement Geometry#buffer for the spherical factory. It's possible I could put in an ad hoc implementation that would work for points (only) since that case isn't too difficult. But unless someone a lot smarter than me can do the non-Euclidean geometry work, I don't see any of the other types getting done any time soon.

As an alternative, you may be able to get away with doing the operation in a projection and then unprojecting the resulting polygon back into geographic coordinates. That would be an approximation, but if the buffer is small and you pick a good projection, it may be good enough for your application.

Regarding distance measures for the simple mercator factory, it returns units in spherical mercator projection coordinates, not in "real world" units. Specifically, the units are meters, scaled by the secant of the latitude (which is the scaling distortion you get in a mercator projection). If you want an approximation of real-world distance units, just multiply the simple mercator distance by the cosine of the latitude (in radians of course) and you'll get meters.

from rgeo.

Interesting. Thank you for the information!

Basically what I was trying to do here was move PostGIS's ST_Buffer for geography over to RGeo (taking points and approximating their circle radius with a polygon), so I could offload some work from the database. Not a big deal though, I can always just fire up a bunch of PostGIS servers for processing if I need to (or implement that method for points only).

from rgeo.

Hmm. That sounds really clumsy. I'll tell you what. I'll reopen this and let me see if I can just get the spherical factory's buffer method implemented for points in the next day or so. I believe the spherical geometry for that case is fairly straightforward.

from rgeo.

Kyle, I pushed rgeo 0.3.13 which includes an implementation of the buffer method for points in the spherical factory (and the simple cartesian factory).

from rgeo.