contain()

[Larson-Li]

As we know, the point is float type, sometime will cause a problem, [(1.0, 1.0), (2.0, 1.0), (2.0, 2.0)] actually is [(1.00000000000000001, 1.00000000000001), (2.0, 1.0), (2.0, 2.0)], this polygon contain point (1.0000000000000001, 1.00000000000001), does not contain (0.9999999999999999, 0.99999999999999)

[heitzmann]

If you need precise comparisons, you can snap the vertices to a grid using, for example, numpy.round with appropriate number of decimals after the transformations you are applying. Alternatively, add some tolerance to the check.

[Larson-Li]

transformation is the way that I used now, in order to accomplish it, I get points of shape, then tranformation and create polygon again, however, there are billions of shapes, this operation costs so much time with python api. If this operation can be integrated into c environment, it will faster.

[Larson-Li]

or when we parse the gds file, we process all points to (precision / unit), it can also reduce memory.