Programmatically configuring viewers

Viewers in Glue are designed to be easily configured with Python. As much as possible, viewer settings are controlled by simple properties on the state attribute of data viewer objects. For example:

import numpy as np

from glue.core import Data, DataCollection
from import GlueApplication
from glue.viewers.scatter.qt import ScatterViewer

# create some data
d = Data(x=np.random.random(100), y=np.random.random(100))
dc = DataCollection([d])

# create a GUI session
ga = GlueApplication(dc)

# plot x vs y, flip the x axis, log-scale y axis
scatter = ga.new_data_viewer(ScatterViewer)

# Modify viewer-level options
scatter.state.x_att =['x']
scatter.state.y_att =['y']
scatter.state.y_log = True

# Modify settings for the (only) layer shown
scatter.state.layers[0].color = 'blue'

# show the GUI

Viewer Options

The state attribute for each viewer is an instance of a viewer state class. Each viewer state object then has a layers attribute that can be used to control individual layers in the viewer (as shown above).

The following table lists for each built-in viewer the classes defining the state for each viewer/layer type. By clicking on the name of the class, you will access a page from the API documentation which will list the available attributes.

Viewer Viewer state Data layer state Subset layer state
ScatterViewer ScatterViewerState ScatterLayerState ScatterLayerState
ImageViewer ImageViewerState ImageLayerState ImageSubsetLayerState
HistogramViewer HistogramViewerState HistogramLayerState HistogramLayerState

Customizing Plots with Matplotlib

If you want, you can directly manipulate the Matplotlib plot objects that underly Glue viewers. This can be useful if you want to create static plots with custom annotation, styles, etc.

From the GUI

Open the IPython terminal window. The application.viewers variable is a list of lists of all the open viewer windows. Each inner list contains the data viewers open on a single tab. Every viewer has an axes attribute, which points to a Matplotlib Axes object:

viewer = application.viewers[0][0]
ax = viewer.axes
ax.set_title('Custom title')
ax.figure.canvas.draw()  # update the plot

From a script

Save the current glue session via File->Save Session. You can reload this session programmatically as follows:

from import GlueApplication
app = GlueApplication.restore('output.glu', show=False)
viewer = app.viewers[0][0]
ax = viewer.axes