Use of Numpy

Numpy is an extension library for the Python programming language which offers support for large, multidimensional arrays and matrices and a number of linear algebra operations that can be performed on them. It's effectively Python's standard go-to library for linear algebra, and is used heavily by other Python libraries in the scientific computing and data analysis spaces.

The underlying engeom Rust library uses nalgebra for dense linear algebra operations and faer for sparse matrix operations. To provide a consistent interface for the Python bindings, the Python engeom library assumes that many users will already be using numpy for bulk operations on arrays of points, vectors, face lists, etc, and so it makes sense to map many of the engeom features which would operate on points/vectors to operate seamlessly on numpy.ndarray objects.

To that end, there are a number of conventions used in engeom regarding the use of numpy arrays:

1. When mapping large arrays of points/vectors, numpy.ndarray objects should be sized as (n, 2) or (n, 3) for 2D and 3D points/vectors, respectively. Rows represent individual points/vectors, and columns represent the x, y, and z, components of the points/vectors.

2. Some engeom objects will provide access to their data as read-only numpy.ndarray objects. For instance, the Mesh object will expose its vertex and face data as numpy.ndarrays. The shaping of these arrays will be consistent with the above convention, and floating point data will be 64-bit.

3. Where index-based lists, such as the face vertices lists of the Mesh objects are used, these will typically be expected to be unsigned integer types. Internally, most are stored as Rust's u32 or usize types. If you get an error while constructing objects, such as the Mesh, make sure that the data you are passing in is of the correct type.