Ram Sriharsha
Ram Sriharsha
The Python API was broken after 1.0.4 release (incompatibility with Spark SQL after Spark 1.5) and we haven't fixed it yet.
It is fairly easy to expose a SQL Front End to Magellan. We should support all the PostGIS geospatial extensions to SQL (the read side)
The ZOrderCurve Index computed by Magellan today subdivides space into equal sized grids of a given precision. This can easily be changed to compute curves of a precision between min,...
Magellan needs documentation that is beginner friendly yet thorough with some details on the implementation as well so developers can start contributing to it. We also need a developer page
Today Magellan supports - [x] Point within Polygon - [x] Polygon contains Point - [x] Polygon intersects Line as first class citizens Among these, the first two use spatial indices...
Apart from Inner Joins, also support Left Outer/ Right Outer Joins using Spatial Indices. eg. ```points.join(polygons, $"point" within $"polygon", "leftouter | rightouter")``` should also leverage spatial indices.
Magellan currently works with Dataframes but not Datasets. This issue addresses the gap by providing efficient encoders to go between Dataframes and Datasets.
Currently we support intersection and Shape assembly as the only supported geospatial operators. Other common ones include: Union, Distance, Intersection, Symmetric Difference, Convex Hull, Envelope, Buffer, Simplify, Polygon Assembly, Valid,...
Currently we support intersects, within and contains (though contains is supported by a weird >? operator since contains is a keyword) Other common predicates include: Touches, Disjoint, Crosses, Overlaps, Equals,...