itrm 1.1.2

Interactive Terminal Utilities

Interactive Terminal Utilities

import itrm

This library offers several functions for visualizing data within the terminal. This project does not exist just because it is cool. It exists because it fills some needs which few other tools do. For many developers, engineers, and scientists, the terminal is where much of their time is spent. Having to switch contexts every time a plot is generated can be time consuming and annoying. Furthermore, most plotting tools have fairly limited analysis capabilities. They are great for generating final, beautiful figures, but not great at quickly inspecting and understanding the data. Also, if you are working with a remote server through SSH, visualizing the data with conventional tools can be very tedious: save the data to a file, transfer the data to a local machine, write a script just to read and plot the data, plot the data, repeat. This library lets you directly visualize and interact with the data, skipping all the tedium.

Configuration

This library will generate a config.ini file in the same directory as the library's installation. You can manually modify the settings in that file. They are

Setting	Default	Description
uni	True	flag to use Unicode characters
ar	0.48	aspect ratio of characters
cmap	4bit	color map
bold	True	flag to use bold plot characters

The first time itrm is used, a configuration file will be generated and its location will be printed to the terminal. Subsequently, you can get the location of the configuration file by

print(itrm.Config.where())

Unicode

Much of the plotting in the terminal performed by itrm relies on Unicode characters. However, properly displaying those characters requires having a monospace font with those specific glyphs defined. In fact, the default plotting mode relies on braille characters, and relatively few fonts define those. If you are looking for a good terminal font which supports all the Unicode used by this library, try out JuliaMono. However, you might not be interested in downloading fonts, so this library can also forego all Unicode characters and only rely on ASCII characters. This is the purpose of the uni setting in the config.ini file.

Aspect Ratio

Because all the plotting by this library uses text, the aspect ratio (ratio of width to height) of the characters affects the apparent aspect ratio of curves. So, a circle might look perfectly round or squashed depending on the font chosen. This does not mean you need a new font, you just need to adjust the aspect ratio setting, ar, in the config.ini file.

Color Map

By default, the color map used is 4bit. Setting the cmap parameter, you can pick any of the following color maps: spectrum, viridis, grays, reds, greens, blues, or 4bit. All but the last color map use platform-independent, 8-bit colors. The last color map, 4bit, lets you control the colors with your terminal settings instead.

Interactive Plots

itrm.iplot(x, y=None, label=None, rows=1, cols=1,
        lg=None, overlay=False, cmap=None):

This function will create an interactive plot with a cursor. The cursor is used to pan, zoom, and identify values in the plot.

Parameters

The iplot function will render all the data points defined by x and y to the terminal. The inputs x and y can be vectors, matrices, or lists of such arrays. Each row of a matrix is treated as a separate curve. Note, this is different from MatPlotLib, in which each column is treated as a separate row. (This difference is intentional, as in the author's opinion varying time along columns means each column in a matrix can be treated as a vector. This arrangement works very well in linear algebra, especially matrix multiplication with a "set" of vectors over time.)

The shapes of x and y do not have to be the same, but they must be compatible. So, x could be a vector and y could be a matrix as long as the length of x equals the number of columns of y.

If only x is given, it will be interpreted as the y values, and the x values will be an array of indices equal in length to y.

If a label is given, this will be printed in the bottom right of the plot box. It can also be a list of strings. If the length of the list is the same as the number of data sets (each row of a matrix is a different data set), then each string in the list will be displayed with the respective data set. If the length of the list is one greater, the first string will be displayed for the whole plot.

The rows and cols parameters let you specify the number of terminal text rows and columns to use for the plot, respectively. For each of these, if the value is less than or equal to 1, it represents a portion of the available space to use. For example, if rows is 0.5, then half the number of rows of the current terminal window will be used for the plot. If the value is greater than 1, it represents the absolute number of rows or columns to use. Also, if the size of the current terminal cannot be obtained, the available space will default to predefined, fixed values.

You can set the x or y axes to logarithmic scaling by setting the lg parameter to one of "x", "y", or "xy". Note that the values reported for the view and the cursor will be in the original scaling, not logarithmic.

To prevent your terminal history from extending each time a new plot is rendered, you can print a new plot over a previous plot by setting the overlay parameter to True. This can be especially useful when there are multiple plots to render (like for an animation) but you do not want your terminal history to fill up quickly.

For an individual plot, you can override the color map defined in the config.ini file by setting the cmap parameter.

Information Bar

At the bottom of the plot, as part of the border, various information sets are listed:

• View
  • X: range of the x axis within view
  • Y: range of the y axis within view
• Cursor
  • n: index of the value the cursor is currently centered on. This is visible only when a specific data set is selected.
  • x: x-axis value corresponding to the current location of the cursor.
  • y: y-axis value or range of values corresponding tot he current location of the cursor.
• Metrics (only visible when a ghost cursor exists)
  • δn: difference of indices between ghost cursor and current cursor. This is visible only when a specific data set is selected.
  • δx: difference of x values between ghost cursor and current cursor.
  • δy: difference of y values between ghost cursor and current cursor. This is visible only when a specific data set is selected.
  • μ: mean of y values from ghost cursor to current cursor. This is visible only when a specific data set is selected.
  • σ: standard deviation of y values from ghost cursor to current cursor. This is visible only when a specific data set is selected.
• Label

When not all of these information sets will fit in the currently available width of the plot, they will be separated into different groups. You can cycle through these groups by pressing the m or M key.

Keybindings

The iplot function provides interactivity through a vertical cursor. You can move the cursor left and right, at normal speed or fast speed. You can zoom in and out. And, you can cycle through which rows of the x and y data to focus on. Note, iplot is designed for monotonically-increasing x values, and, consequently, does not support equal axis scaling.

The following table details the shortcut keys:

Keys	Function		Keys	Function
q, ⌫, ↵	exit interactive plot		j, s, ↓	zoom in
h, a, ←	move cursor left		k, w, ↑	zoom out
l, d, →	move cursor right		J, S, ⇧↓	zoom in fast
H, A, ⇧←	move cursor left fast		K, W, ⇧↑	zoom out fast
L, D, ⇧→	move cursor right fast		n	select next data set
g	move cursor to start		N	select previous data set
G	move cursor to end		m	next info set
c, z	center view on cursor		M	previous info set
x	toggle x log scaling		v	toggle ghost cursor
y	toggle y log scaling		f	start function
i	toggle individual view		F	restore original data

Note that in Windows terminal emulators, there is no support for shift-arrow keys. Instead, use alt-arrow keys.

Individual Data Sets

When many data sets are being plotted simultaneously, it can be helpful to hide all other data sets with the i key in order to just see the selected data set.

Ghost Cursor

If you want to make a comparison between two points, you can use the ghost cursor. First, position the cursor at the start position. Then, press the v key. Immediately, you should see in the bottom left corner, several metrics (only the δx metric if the cursor is white and there are multiple data sets). These metrics are detailed in the Information Bar section above. Moving the cursor will leave behind a ghost. As the cursor moves, the metrics will update to reflect the range of values from the ghost cursor to the current cursor.

Functions

Without writing any code, you can run a number of functions on the data and see the results. First, press the f key. Then, follow that with other keys to get the specific function applied to the data. The following table shows the full key sequences:

Keys	Description
fd	Derivative of y with respect to x
fi	Integral of y with respect to x
ff	Magnitude of Fourier transform of y
ftl	Trim (remove) data left of cursor
ftr	Trim (remove) data right of cursor
f#a	Weighted moving average of half width #
f#d	De-trend data with polynomial of degree #

(Other functions are planned for the future.) When one of the functions have been applied, the sides of the plotting box will be rendered gray. You can restore the original data by pressing the F key.

Plots

itrm.plot(x, y=None, label=None, rows=1, cols=1,
        ea=0, lg=None, overlay=False, cmap=None):

The plot function is a non-interactive version of the iplot function. All of the same parameters are provided, with the addition of the equal-axes (ea) parameter, which enables plotting things like circles without them rendering as ellipses. This function does not require monotonicity of the x-axis data.

Single curve	Multiple curves

Bars

itrm.bars(x, labels=None, cols=1)

It can be convenient to plot a simple bar graph. The x input is the vector of values. The labels input is a list of strings corresponding to the labels to print before the bar of each value in x. If the cols input is greater than 1, it is the total width of characters including the labels. If it is less than or equal to 1, it is the portion of the terminal window width which will be used for the graph. If the fat input is set to True, the bars will be thick.

Heat maps

itrm.heat(matrix)

The heat function will generate a heat map of the matrix input using 24 shades of gray. Black is used for the lowest value and white for the highest value. If itrm.config.uni is True, half-block characters from the Unicode table will be used. If it is False, two spaces per element of the matrix will be used.

With Unicode	Without Unicode

Tables

itrm.table(matrix, head=None, left=None, width=10, sep='  ')

You can print a nicely spaced table of the matrix data. The head and left inputs are lists of header and left-most column labels, respectively, to print around the matrix.

Sparsity

itrm.spy(matrix)

If all you want to see is the sparsity of a matrix, use this function.

With Unicode	Without Unicode

Progress bars

bar = itrm.Progress(K, cols=1)
bar.update(k)

There are many progress bar libraries available for Python. But, many of them seem to be extremely over-complicated. TQDM, for example, includes over 20 source files. This library's implementation of a progress bar is a single, one-page function. The k input is the counter of whatever for loop the progress bar is reporting on. The K input is one greater than the largest possible value of k, as in for k in range(K):. When the process is completed, the total elapsed time will be displayed. If cols is not provided, the full width of the current terminal window will be used.