Is there a context manager for temporarily changing matplotlib settings?

Question

In pandas and seaborn, it is possible to temporarily change the display/plotting options by using the with keyword, which applies the specified setting only to the indented code, while leaving the global settings untouched:

print(pd.get_option("display.max_rows"))

with pd.option_context("display.max_rows",10):
    print(pd.get_option("display.max_rows"))

print(pd.get_option("display.max_rows"))

Out:

60
10
60

When I similarly try with mpl.rcdefaults(): or with mpl.rc('lines', linewidth=2, color='r'):, I receive AttributeError: __exit__.

Is there a way to temporarily change the rcParams in matplotlib, so that they only apply to a selected subset of the code, or do I have to keep switching back and forth manually?

Answer 1

Yes, using stylesheets.

See: http://matplotlib.org/users/style_sheets.html

e.g.:

# The default parameters in Matplotlib
with plt.style.context('classic'):
    plt.plot([1, 2, 3, 4])

# Similar to ggplot from R
with plt.style.context('ggplot'):
    plt.plot([1, 2, 3, 4])

You can easily define your own stylesheets and use

with plt.style.context('/path/to/stylesheet'):
    plt.plot([1, 2, 3, 4])

For single options, there is also plt.rc_context

with plt.rc_context({'lines.linewidth': 5}):
    plt.plot([1, 2, 3, 4])

Answer 2

Yes, the matplotlib.rc_context function will do what you want:

import matplotlib as mpl
import matplotlib.pyplot as plt
with mpl.rc_context({"lines.linewidth": 2, "lines.color": "r"}):
    plt.plot([0, 1])