ggshow

Produce ggplot2 graphs from Python

Install

# from pypi
pip install ggshow

# or from github
git clone --depth 1 https://github.com/kota7/ggshow.git
pip install -U ./ggshow

Requirements

• Python 3.7+ (but would work in some older versions)
• Rscript command and the R environment with ggplot2 package installed

Functionalities

• ggshow: Draw graphs using ggplot2 on R and return the Image object that you can show on Jupyter notebook. From v0.1.10, we can also save the image to the file.
• ggwrite: Draw graphs using ggplot2 on R and save it to a file (but not show on the notebook).
• %gg, %%gg: ipython magic for drawing ggplot2 graph.

from ggshow import ggshow, ggwrite

# Minimal example
# pass R code to produce ggplot graph to ggshow function
# note: ggplot2 is automatically imported
ggshow("""
  x <- c(1,2,3)
  y <- c(4,5,6)
  ggplot(mapping=aes(x, y)) + geom_line()
""", savesize=(3, 2))
None

# Example with data frames
# pass pandas data frames as keyword arguments
# then you can use them in the R code
import numpy as np
import pandas as pd
x = np.linspace(-20, 20, 500)
y = np.sin(x)
y2 = 0.4*np.cos(x)

df1 = pd.DataFrame({"x":x, "y":y})
df2 = pd.DataFrame({"x":x, "y":y2})
ggshow("""
  ggplot(a, aes(x, y)) + 
    geom_line(color="blue") +
    geom_line(data=b, linetype="dashed", color="red") +
    theme_bw()
""", dispwidth=500, savesize=(4, 2), a=df1, b=df2)
None

# Example to save the graph to a file
ggshow("""
  x <- c(1,2,3)
  y <- c(4,5,6)
  ggplot(mapping=aes(x, y)) + geom_line()
""", outfile="foo.jpg", savesize=(3, 2))

# Example to "only" save the graph to a file, not to show
ggshow("""
  x <- c(1,2,3)
  y <- c(4,5,6)
  ggplot(mapping=aes(x, y)) + geom_line()
""", outfile="bar.jpg", savesize=(3, 2))

# Confirm the file has been created
from IPython.display import Image
Image("bar.jpg", width=400)

# Example to specify the image file format
# But there may not be significant differences to human eyes
# Default: png
ggshow("""
  x <- c(1,2,3)
  y <- c(4,5,6)
  ggplot(mapping=aes(x, y)) + geom_line() + ggtitle("PNG")
""", savesize=(3, 2), imageformat="png")

ggshow("""
  x <- c(1,2,3)
  y <- c(4,5,6)
  ggplot(mapping=aes(x, y)) + geom_line() + ggtitle("JPEG")
""", savesize=(3, 2), imageformat="jpeg")

ggshow("""
  x <- c(1,2,3)
  y <- c(4,5,6)
  ggplot(mapping=aes(x, y)) + geom_line() + ggtitle("SVG")
""", savesize=(3, 2), imageformat="svg")
None

# Example using %gg magic
%load_ext ggshow

## line magic
%gg ggplot(mapping=aes(1, 2)) + geom_point() -s 4 3 --dispwidth 300 --outfile buz.jpeg

## line magic with the code as a string variable
x = np.linspace(-20, 20, 500)
y = np.sin(x)
y2 = 0.4*np.cos(x)

df1 = pd.DataFrame({"x":x, "y":y})
df2 = pd.DataFrame({"x":x, "y":y2})
code = """
ggplot(a, aes(x, y)) + 
  geom_line(color="blue") +
  geom_line(data=b, linetype="dashed", color="red") +
  theme_bw() +
  ggtitle("Sine and Cosine Waves")
"""

%gg {code} -s 4 2 --dispwidth 400 --data a=df1 b=df2

%%gg -s 4 2 --dispwidth 400 --data a=df1 b=df2

## cell magic
ggplot(a, aes(x, y)) + 
  geom_line(color="blue") +
  geom_line(data=b, linetype="dashed", color="red") +
  theme_bw() +
  ggtitle("Sine and Cosine Waves")

# show the full command options
%gg --help

Help on method gg in module ggshow.ggshow:

gg(line, cell=None) method of ggshow.ggshow.GGMagic instance
    ::

      %gg [--help] [--outfile OUTFILE] [-s SAVESIZE SAVESIZE] [--scale SCALE]
              [--units UNITS] [--dpi DPI] [--message_encoding MESSAGE_ENCODING]
              [--rscriptcommand RSCRIPTCOMMAND] [-w DISPWIDTH] [-h DISPHEIGHT]
              [--imageformat {png,jpeg,svg}] [--libs [LIBS ...]]
              [--data [DATA ...]]
              [plotcode ...]

    positional arguments:
      plotcode              R code

    options:
      --help
      --outfile OUTFILE     File path to save the graph
      -s <SAVESIZE SAVESIZE>, --savesize <SAVESIZE SAVESIZE>
                            height, width
      --scale SCALE         ggsave option scale
      --units UNITS         ggsave option units
      --dpi DPI             ggsave option dpi
      --message_encoding MESSAGE_ENCODING
                            Encoding of the subprocess outputs
      --rscriptcommand RSCRIPTCOMMAND
                            Rscript command name
      -w DISPWIDTH, --dispwidth DISPWIDTH
                            display width
      -h DISPHEIGHT, --dispheight DISPHEIGHT
                            display width
      --imageformat <{png,jpeg,svg}>
                            imagefile format
      --libs <[LIBS ...]>   R libraries to use
      --data <[DATA ...]>   data frames mapping as {name in r}={name in python}

# By default, we use 'Rscript' as the command to run R code.
# If this is not a valid command on the environment,
# we can specify the command by `rscriptcommand` option 
ggshow("""
  x <- c(1,2,3)
  y <- c(4,5,6)
  ggplot(mapping=aes(x, y)) + geom_line()
""", savesize=(3, 2), rscriptcommand="/usr/local/bin/Rscript")


# Or set the new command path as the default using set_rscript function.
from ggshow import config, set_rscript

set_rscript("/usr/local/bin/Rscript")  
print(config.rscript)
# this is just the full path of the command on this environment
# so the command will work in the same way

/usr/local/bin/Rscript