The Mapper output window

Here is a typical Mapper output from one of the toy examples (the horse):

General navigation

See here: Navigation in the figure windows

Node highlighting

• Click on a node to highlight it.

• Shift-Click allows to select and de-select multiple nodes.

  

  Mapper output window with several highlighted nodes.

• Click on the white background to reset the selection.

• Keyboard navigation

  • L selects all nodes in the same level set(s) as the currently highlighted node(s).
  • < moves to the next lower level set.
  • > moves to the next higher level set.

• When 2- or 3-dimensional data is displayed, the data points in the selected nodes are highlighted accordingly.

  

  Data points in the selected nodes above are highlighted.

• The identities of the selected data points can be saved to a file, see Save highlighted nodes below.

Menus

“File” menu

Save figure Ctrl-S

Save the figure in the current view. A variety of file formats is available, for both vector and bitmap graphics.

Save highlighted nodes Ctrl-Shift-S

In order to check which actual data points are represented by a set of nodes, the list of data points in the currently highlighted nodes can be saved to a text file. The text file is a simple list of indices (one number per line, separator: newline, LF). The first data point has index 0.

“Options” menu

Relabel Alt-L

(Advanced.) This allows to display custom node labels. A file dialog asks for a Python script with a function that determines the labels from the node data. The Python script must define the following objects:

name

a string with the name of the labelling scheme (for the figure legend).

label

a function that is given a node (ie. a mapper.mapper_output.node object) and must return a string with the node label.

In short, a node object contains the following data:

level

level identifier, usually an integer or a tuple of integers

points

a numpy.ndarray(N, dtype=int); these are the indices of the points in the data set which are contained in the node.

attribute

any attribute, eg. a color, normally the average filter value in the node.

Below is a small example which replaces the standard label (node size) with the average filter value.

name = 'filter value'

def label(node):
    return '{:.3n}'.format(node.attribute)

“View” menu

Reset 1

Restore the original view.

Show labels Ctrl-L

Show or hide the node labels. The default node labels are nodes sizes (ie. number of data points in each node).

Mapper output with and without labels.

Resize window Ctrl-R

Resize the figure window to a given exact size in pixels. Enter eg. 1000x500 for a canvas of size 1000×500, or just 1000 for width 1000 and height corresponding to the figure aspect ratio.