Series of minor tweaks and a lot of documentation polishing, more examples and a few plots now.

New in 1.1.8

Instead of writing

 R = SO3(np.array([[0, 1, 0], [1, 0, 0], [0, 0, -1]]))

you can now write

 R = SO3([[0, 1, 0], [1, 0, 0], [0, 0, -1]])

works for SO2, SO3, SE2, SE3.

There is now a .eulervec method for SO3 and SE3 classes, returns a 3-vector of exponential coordinates.

The .prod method could produce a denormalised result given a long sequence of transforms, and extra options now control how to deal with that.

New since 1.1.0

Bug fixes

transform interpolation, converting rotation matrix to unit quaternion,

New features

Some fixes to color/styling of plots.

A .conjugation method

A = np.array([[10, 0], [0, 1]])
AA = R.conjugation(A)) # is equivalent to R * A * R.inv()

Plot functions accept a boolean argument new which will render the graphics in a new plot, not the current plot.

Animate function can return an HTML5 video, useful for rendering in notebooks.

Added function to base that converts a MATLAB like matrix as a string to an ndarray

str2array("5")
str2array("[1 2 3]")
str2array("[1 2; 3 4]")
str2array(" [ 1 , 2 ; 3 4 ] ")
str2array("[1; 2; 3]")

The Ellipse class is more complete.

Type hinting support