tabplot

WORK IN PROGRESS

When working from the terminal, I'd need to quickly (and beautifully) plot the data in csv-style text files. This meant good colors, axis labels, plot labels, adjustable axis limits, etc. I started with making a simple wrapper script around matplotlib to do this. Over time, I kept adding pre- and post-processing functionality such as scaling, normalizing, and linear regression, not to mention exposing more of matplotlib's properties and functions. Since the script kept growing, I decided to move it into its own separate package.

Installation

You can install tabplot via testpypi

pip install git+https://github.com/jayghoshter/tabplot

Check out SciencePlots for added plot beautification.

Examples

Simple line plots

Assuming text data stored in csv-style files with the following assumptions:

• No header labels
• x,y data in columns 0,1 respectively
• Either space or comma separated

The following code will generate basic plots. For customization options please refer to the examples section or test cases.

from tabplot import Plot
plot = Plot()
plot.files = ['./file_1.csv', './file_2.csv'],
plot.linewidths = 2,
plot.labels = ['A', 'B']
plot.read()
plot.draw()
plot.show()
plot.close()

or

from tabplot import Plot
Plot(
    files = ['./file_1.csv', './file_2.csv'],
    linewidths = 2,
    labels = ['A', 'B']
).read().draw().show().close()

• Most used plot properties are exposed directly on the Plot object.
• files, twinx, and labels can be specified on Plot or in the read() method
• read() loads the data from the given files into .xs and .ys, which can be accessed and manipulated as wished.
• setup() creates a .fig and .ax if not already done
• draw() plots the lines etc. to the figure
• save('file.pdf') will write to file.pdf
• show() will display the plotted image

Bar plots

from tabplot import BarPlot
files = ['dummy.csv']
BarPlot(bar_width = 0.5).read(files).draw().save('dummy.pdf').close()

• BarPlot is derived from Plot

Notes/Issues

• Operator chaining is possible as seen above.
• Currently only uses one subplot
• Make sure to close the current plot with .close(). Unless garbage collected, plots may linger and calling .show() for some other object may also display a previous unclosed figure.
• hlines/vlines must be called after all plots/fit_lines etc are done so that no further changes to xlims, ylims are effected later.
• legends don't include hlines
• preload-style can affect legend style even if not explicitly specified because we use setrc() in _update_params. So the default value overrides anything given in the style. While this is to be expected, the question is if this is user-friendly or not. Maybe a n otice is sufficient, that --preload-style can have this behavior.
  • [TASK] perhaps I can separate out the setrc_defaults() and call it at another time? Or only call it if the user has specified changes to the default values