Module reference

If the GUI does not offer enough flexibility, you may always write your own Python code. Example scripts can be produced with the GUI menu item File→Generate standalone Python script. The module reference below documents the relevant functions and data structures in the mapper module. (To be completed...!)

mapper.mapper(pcd, filt, cover, cutoff, mask=None, cluster=<mapper._mapper.single_linkage instance>, point_labels=None, metricpar={}, simple=False, filter_info=None, verbose=True)

Mapper algorithm

Parameters:

• pcd (numpy.ndarray((N,n), dtype=float) or numpy.ndarray((N*(N-1)/2), dtype=float)) – input data, point cloud in \(R^n\), or compressed distance matrix for N points
• filt (numpy.ndarray((N, comp), dtype=float) or numpy.ndarray(N, dtype=float)) – filter function with comp components
• cover (iterator) – Class for the cover of the filter range. See Cover methods.
• cutoff (function or None) – Cutoff function for the partial clustering tree. See section:cluster_cutoff.
• mask (Anything that can be used for indexing of NumPy arrays, e.g. a Boolean array of size N.) – (Mainly for the GUI) A mask to choose a subset of the input points
• cluster (See section:clustering_function) – Clustering function.
• point_labels (numpy.ndarray(N)) – Labels for the input points (optional). If this is None, the points are labeled 0,1,...,N−1.
• metricpar (dict) – If the input data is in vector form, these are the parameters that are given to the scipy.spatial.distance.pdist function. If the input data is a compressed distance matrix, this argument is ignored.
• simple (bool) – (to be documented by example) If True, then intersections are only considered for adjacent cover patches in the 1-dimensional variant. In particular, the output simplicial complex is a graph without higher-dimensional simplices.
• filter_info – (For the GUI) Filter info to be stored in the output
• verbose (bool) – Print status message?

Returns:

Mapper output data structure

Return type:

mapper_output instance