-
Notifications
You must be signed in to change notification settings - Fork 1.2k
Open
Description
Is your feature request related to a problem? Please describe.
Nearly 100% of the time when I generate heatmaps, the x and y axes do not have integer values. Instead, they are equispaced lists defined by the scenario, and it is annoying to infer the list value from the figure pixel index when the figure doesn't include appropriately scaled axes.
Describe the solution you'd like
I would like to see two optional arguments to visdom's heatmap(), specifically, I'm hoping the user would be provide optional inputs like
'x':horizontal_axis_vector
'y':vertical_axis_vector So that an example use would be
import visdom
import numpy as np
viz = Visdom()
horizontal_axis_vector= np.arange(-5,6)[np.newaxis]
vertical_axis_vector = (np.arange(0,10)*.2)[np.newaxis]
h = np.square(vertical_axis_vector.T @ horizontal_axis_vector)
viz.heatmap(h, opts={'x':horizontal_axis_vector, 'y':vertical_axis_vector}) Describe alternatives you've considered
The input-output behavior would be similar to matplotlib's imshow() or matlab's imagesc()
Metadata
Metadata
Assignees
Labels
No labels