savefigs
Easily save all open Matplotlib figures, with useful filenames.
Usage
Assume we have a script some_script.py
that creates multiple Matplotlib figures.
Import the savefigs
function:
from savefigs import savefigs
The below examples assume the figures do not have labels (fig.get_label()
).
If a figure does have a label, it will be used in place of fig{num}
.
Default save settings (./{script filename stem}{figure label or fig{num}}.png
):
savefigs()
# ./some_script_fig1.png, ./some_script_fig2.png, ...
Specify directory:
savefigs(save_dir="figs") # must exist
# ./figs/some_script_fig1.png, ./figs/some_script_fig2.png, ...
Specify a different prefix to the base stem format:
savefigs(stem_prefix="run1")
# ./run1_fig1.png, ./run1_fig2.png, ...
Save in multiple file formats:
savefigs(formats=["png", "pdf"])
# ./some_script_fig1.png, ./some_script_fig1.pdf, ...
Avoid overwriting files:
savefigs(clobber=False, noclobber_method="add_num")
# ./some_script_fig3.png (assuming ./some_script_fig{1,2}.png already exist)
noclobber_method
), setting clobber=False
will instead error.
Background
When writing a script that creates multiple figures, I usually label them (usually using the num
argument to plt.figure()
/plt.subplots()
), which makes it easier to find the correct figure window. Then, at the end of the script I write a loop like:
for num in plt.get_fignums():
fig = plt.figure(num)
fig.savefig(f"{fig.get_label()}.pdf", ...)
# Maybe another format...
savefigs()
essentially does this, but is more robust and provides additional features through keyword arguments. And it saves having to write those lines in the script, instead allowing the simple one-liner:
from savefigs import savefigs; savefigs()