teeplot wrangles your data visualizations out of notebooks for you
- Free software: MIT license
- Installation:
python3 -m pip install teeplot
- Documentation: https://github.com/mmore500/teeplot/blob/master/README.rst
- Repository: https://github.com/mmore500/teeplot
teeplot's tee
function can wrap your plotting calls to automatically manage matplotlib file output, picking meaningful names based on semantic plotting variables.
# adapted from https://seaborn.pydata.org/generated/seaborn.FacetGrid.html#seaborn.FacetGrid
import seaborn as sns; from teeplot import teeplot as tp
tp.tee(sns.lmplot, # plotter
sns.load_dataset("tips"), col="time", hue="sex", x="total_bill", y="tip", # fwded kw/args
teeplot_postprocess=sns.FacetGrid.add_legend) # teeplot options
teepots/col=time+hue=sex+post=add_legend+viz=lmplot+x=total-bill+y=tip+ext=.pdf teepots/col=time+hue=sex+post=add_legend+viz=lmplot+x=total-bill+y=tip+ext=.png
Here's how it works: teeplot's tee
function that acts as a wrapper around your plotting calls.
Give tee
your plotting function (e.g., sns.lineplot
) as the first argument and then add the arguments you want to call it with.
teeplot automatically captures the function and its arguments, calls the plotter as instructed, and then it handles the matplotlib file output for you. It generates descriptive filenames for the saved plots by extracting key information from the plot parameters and arguments. This feature allows you to keep track of your visualizations easily by making the process of saving and cataloging your plots more efficient, systematic and meaningful, taking the hassle out of manual file management.
teeplot contains several advanced features, such as a draftmode
flag, which will disable file output globally, and the teeplot_callback
kwarg, which delays plot output to allow for figure tweaks.
Read on for details.
Contents
- Usage : Example 1 | Example 2 | Example 3 | Example 4 | Example 5
- API : teeplot.tee() | Module-Level Configuration | Environment Variables
- Credits
Usage
Example 1
Simple example demonstrating use with pandas built-in plotting.
# adapted from https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.plot.box.html
import pandas as pd; from teeplot import teeplot as tp
age_list = [8, 10, 12, 14, 72, 74, 76, 78, 20, 25, 30, 35, 60, 85]
df = pd.DataFrame({"gender": list("MMMMMMMMFFFFFF"), "age": age_list})
tp.tee(df.plot.box, # plotter...
column="age", by="gender", figsize=(4, 3)) # ...forwarded kwargs
teepots/by=gender+column=age+viz=box+ext=.pdf teepots/by=gender+column=age+viz=box+ext=.png
Example 2
Example with seaborn showing use of teeplot_callback
kwarg to allow for plot tweaks before saving.
# adapted from https://seaborn.pydata.org/examples/horizontal_boxplot.html
from matplotlib import pyplot as plt
import seaborn as sns
from teeplot import teeplot as tp
saveit, ax = tp.tee( # create a callback object to finalize plot
sns.boxplot, # plotter...
sns.load_dataset("planets"), # ...forwarded arg & kwargs
x="distance", y="method", hue="method", palette="vlag",
whis=[0, 100], width=.6, # ... and then teeplot options
teeplot_callback=True, teeplot_postprocess="teed.set_xscale('log')")
ax.xaxis.grid(True) # now some tweaks
ax.set(ylabel="")
sns.despine()
plt.gcf().set_size_inches(10, 4)
saveit() # dispatch output callback
teepots/hue=method+palette=vlag+post=teed-set-xscale-log+viz=boxplot+x=distance+y=method+ext=.pdf teepots/hue=method+palette=vlag+post=teed-set-xscale-log+viz=boxplot+x=distance+y=method+ext=.png
Example 3
Example with matplotlib, also showing use of teeplot_callback
kwarg .
# adapted from https://matplotlib.org/stable/tutorials/pyplot.html
from matplotlib import pyplot as plt
import numpy as np; from teeplot import teeplot as tp
data = {'a': np.arange(50), 'c': np.random.randint(0, 50, 50),
'd': np.random.randn(50)}
data['b'], data['d'] = data['a'] + 10 * np.random.randn(50), np.abs(data['d']) * 100
saveit, __ = tp.tee( # create a callback object to finalize plot
plt.scatter, # plotter...
data=data, x='a', y='b', c='c', s='d', # ...forwarded kwargs
teeplot_callback=True) # teeplot options
plt.xlabel('entry a') # now some tweaks
plt.ylabel('entry b')
plt.gcf().set_size_inches(5, 3)
saveit() # dispatch output callback
teepots/c=c+s=d+viz=scatter+x=a+y=b+ext=.pdf teepots/c=c+s=d+viz=scatter+x=a+y=b+ext=.png
Example 4
Example with seaborn FacetGrid
demonstrating use of exec
'ed teeplot_postprocess
that adds a map_dataframe
step over the teed
result value and also results in additional semantic information being added to plot filenames (under the "post=
" key).
# adapted from https://seaborn.pydata.org/generated/seaborn.FacetGrid.html#seaborn.FacetGrid
import seaborn as sns
from teeplot import teeplot as tp
tp.tee(
sns.FacetGrid, # plotter...
sns.load_dataset("tips"), # ...forwarded args & kwwargs
col="time", hue="sex", aspect=1.5,
teeplot_postprocess="teed.map_dataframe(sns.scatterplot, x='total_bill', y='tip')")
teepots/col=time+hue=sex+post=teed-map-dataframe-sns-scatterplot-x-total-bill-y-tip+viz=facetgrid+ext=.pdf teepots/col=time+hue=sex+post=teed-map-dataframe-sns-scatterplot-x-total-bill-y-tip+viz=facetgrid+ext=.png
Example 5
Demonstration of teeplot use with a custom function.
Note the function name automatically used as "viz=
" key in output filenames.
# adapted from https://seaborn.pydata.org/examples/pairgrid_dotplot.html
import seaborn as sns; from teeplot import teeplot as tp
df = sns.load_dataset("car_crashes")
def dot_plot(data, x_vars, y_vars): # custom plotter
g = sns.PairGrid(data.sort_values("total", ascending=False),
x_vars=x_vars, y_vars=y_vars,
height=5, aspect=0.66)
g.map(sns.stripplot, size=10, orient="h", jitter=False,
palette="flare_r", linewidth=1, edgecolor="w")
for ax in g.axes.flat:
ax.xaxis.grid(False)
ax.yaxis.grid(True)
tp.tee(
dot_plot, # plotter, then forwarded args/kwargs
df[df["abbrev"].str.contains("A")], x_vars=df.columns[:-3], y_vars=["abbrev"],
teeplot_outinclude=["x_vars", "y_vars"], teeplot_save={".eps", ".png"})
teeplots/viz=dot-plot+x-vars=index-total-speeding-alcohol-not-distracted-no-previous-dtype-object+y-vars=abbrev+ext=.eps teeplots/viz=dot-plot+x-vars=index-total-speeding-alcohol-not-distracted-no-previous-dtype-object+y-vars=abbrev+ext=.png
API
teeplot.tee()
Executes a plotting function and saves the resulting plot to specified formats using a descriptive filename automatically generated from plotting function arguments.
Parameter | Description |
---|---|
plotter |
The plotting function to be executed. Required. |
Additional args & kwargs | Forwarded to the plotting function and used to build the output filename. |
teeplot_callback |
If True, returns a tuple with a callback to dispatch plot save instead of immediately saving the plot after running the plotter. Default is False. |
teeplot_dpi |
Resolution for rasterized components of saved plots, default is publication-quality 300 dpi. |
teeplot_oncollision |
Strategy for handling filename collisions: "error", "fix", "ignore", or "warn", default "warn"; inferred from environment if not specified. |
teeplot_outattrs |
Dict with additional key-value attributes to include in the output filename. |
teeplot_outdir |
Base directory for saving plots, default "teeplots". |
teeplot_outinclude |
Attribute keys to always include, if present, in the output filename. |
teeplot_outexclude |
Attribute keys to always exclude, if present, from the output filename. |
teeplot_postprocess |
Actions to perform after plotting but before saving. Can be a string of code to exec or a callable function. If a string, it's executed with access to plt and sns (if installed), and the plotter return value as teed . |
teeplot_save |
File formats to save the plots in. Defaults to global settings if True , all output suppressed if False . Default global setting is {" .png", ".pdf"} . Supported: ".eps", ".png", ".pdf", ".ps", ".svg". |
teeplot_show |
Dictates whether plt.show() should be called after plot is saved. If True, the plot is displayed using plt.show() . Default behavior is to display if an interactive environment is detected (e.g., a notebook). |
teeplot_subdir |
Optionally, subdirectory within the main output directory for plot organization. |
teeplot_transparent |
Option to save the plot with a transparent background, default True. |
teeplot_verbose |
Toggles printing of saved filenames, default True. |
Return Value: returned result from plotter call if teeplot_callback
is False
, otherwise tuple of save-plot callback and result from plotter call.
Module-Level Configuration
-
teeplot.draftmode
: A boolean indicating whether to suppress output to all file formats. -
teeplot.oncollision
: Default strategy for handling filename collisions, with options like 'error', 'fix', 'ignore', or 'warn'. -
teeplot.save
: A dictionary mapping file formats (e.g., ".png") to default save behavior asTrue
(always output),False
(never output), orNone
(defer to call kwargs).
Environment Variables
-
TEEPLOT_ONCOLLISION
: Configures the default collision handling strategy. Seeteeplot_oncollision
kwarg -
TEEPLOT_DRAFTMODE
: If set, enables draft mode globally. -
TEEPLOT_<FORMAT>
: Boolean flags that determine default behavior for each format (e.g.,EPS
,PNG
,PDF
,PS
,SVG
); "defer" defers to call kwargs. -
CI
, etc.: If a continuous integration environment is detected, defaultteeplot_save
behavior will output only.pdf
files, instead of.pdf
and.png
files. This can be overridden withTEEPLOT_<FORMAT>
.
Credits
Output filenames are constructed using the keyname package.
This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.