Bokeh: dashboard not updating axis range based on widget value - python-2.7

Question

I am trying to implement an interactive dashboard in Bokeh with a "play" function that loops through all value pairs for two indicators selected by widgets.
Screen cap of dashboard
While the loop works, the dashboard resets the axis values for each step of the loop. So what I need is to set axis values based on the widget.value selected. To this end, I have built a data frame "ranges" that has the name of the indicator as index and the min/max value for each indicator as columns.
The updates for controls work thusly (x_axis,etc. are the names of the widgets):
controls = [x_axis, y_axis, start_yr, end_yr, years]
for control in controls:
control.on_change('value', lambda attr, old, new: update())
The update function is supposed to update the ranges upon change in the controls like this:
def update():
p.x_range = Range1d(start = ranges.loc[x_axis.value,"Min"],
end = ranges.loc[x_axis.value,"Max"])
p.y_range = Range1d(start = ranges.loc[y_axis.value,"Min"],
end = ranges.loc[y_axis.value,"Max"])
What should happen: Whenever I change the value of the widget, the ranges should update, but other than that, they should remain constant
What does happen: The ranges are set based on the value of the widget initially set and don't change on update.
I've tried to find examples trying to achieve something similar but no luck.

Answer 1

This is a working example:
import numpy as np
from bokeh.plotting import figure
from bokeh.models import Range1d
from bokeh.io import curdoc
x = np.linspace(0, 100, 1000)
y = np.sin(x)
p = figure(x_range=(0, 100))
p.circle(x, y)
def cb():
# this works:
p.x_range.start += 1
p.x_range.end += 1
# this also works:
#p.x_range = Range1d(p.x_range.start+1, p.x_range.end+1)
curdoc().add_periodic_callback(cb, 200)
curdoc().add_root(p)