**Programming language**: Crystal

**Tags**: Science And Data Analysis

**Latest version**: v0.7.1

## ishi alternatives and similar shards

Based on the "Science and Data analysis" category

** Code Quality Rankings and insights are calculated and provided by Lumnify.*

*They vary from L1 to L5 with "L5" being the highest. Visit our partner's website for more details.*

*Do you think we are missing an alternative of ishi or a related project?*

### Popular Comparisons

## README

## 石の上にも三年

Graph plotting package with a small API and sensible defaults powered by gnuplot.

Requires gnuplot.

## Installation

- Add the dependency to your
`shard.yml`

:

```
dependencies:
ishi:
github: toddsundsted/ishi
```

- Run
`shards install`

## Usage

To display a line chart of the data points in `xdata`

(the x values)
and `ydata`

(the corresponding y values):

```
require "ishi"
ishi = Ishi.new
ishi.plot(xdata, ydata)
ishi.show
```

Or, if you prefer command-style syntax:

```
require "ishi"
Ishi.new do
plot(xdata, ydata)
end
```

A chart can display multiple plots. The following code displays two plots in one chart: one derived from discrete data points and the other from the equation of a line.

```
require "ishi"
Ishi.new do
plot([1, 2, 3, 4, 5], [1.0, 1.4, 1.9, 2.4, 2.6], "ko", title: "data")
plot("0.4 * x + 0.7", "b--")
end
```

The etc/examples directory contains examples of usage.

### plot

`plot`

plots points and lines. It takes data in several formats:

`plot(ydata)`

- y values in*ydata*with x values ranging from`0`

to`ydata.size - 1`

`plot(xdata, ydata)`

- x values in*xdata*and corresponding y values in*ydata*`plot(xdata, ydata, zdata)`

- x values in*xdata*, y values in*ydata*, and z values in*zdata*`plot(expression)`

- any gnuplot-supported mathematical expression

*xdata*, *ydata* and *zdata* may be any type that implements
`Indexable(Number)`

. Chart dimensionality (2D or 3D) is inferred from
the data.

All `plot`

methods/commands accept the optional named arguments
*title*, *style*, *dashtype* (*dt*), *linecolor* (*lc*), *linewidth*
(*lw*), *pointsize* (*ps*), *pointtype* (*pt*), *linestyle* (*ls*),
*fillstyle* (*fs*) and *format*.

*title* specifies the title of the plot in the chart key.

*style* explicitly specifies the style of the plot (`:lines`

,
`:points`

, etc.). Ishi will try to infer the style from the data and
the other named arguments (for example, if both *linewidth* and
*pointsize* are provided, the style will be `:linespoints`

).

By default, plots are rendered with solid lines. *dashtype* (*dt*)
specifies the pattern of dashes to use instead. *dashtype* may be an
array of pairs of numbers that specify the length of a solid line
followed by the length of an empty space (e.g. `[2, 3, 5, 7]`

), or it
may be a string composed of . (dot), - (hyphen), _ (underscore) and
(space), which is then converted into an array as follows: each "."
becomes `[2, 5]`

, "-" becomes `[10, 10]`

, "_" becomes `[20, 10]`

and
" " adds 10 to the previous empty value.

*linecolor* specifies the color to use for lines and points. Colors
may be specified by name (e.g. "red", "blue", or "green") or by
hexadecimal color value (e.g. "#AARRGGBB" or "#RRGGBB").

*linewidth* and *pointsize* scale the width of lines and the size of
points, respectively. A value of 2.0 is twice as big as the default
width or size.

The following code demonstrates the use of *dashtype*, *linecolor* and
*linewidth*:

```
require "ishi"
Ishi.new do
plot("x + 4.0", dashtype: "-", linewidth: 2)
plot("x + 3.0", dashtype: "_", linewidth: 2)
plot("x + 2.0", dashtype: ".", linewidth: 2)
plot("x + 1.0", dashtype: "..._", linewidth: 2)
plot("x + 0.0", dashtype: [30, 10, 50, 10], linewidth: 2, linecolor: "#88001100")
end
```

*pointtype* selects the type of point to render. Available types
depend on the gnuplot terminal device used, but commonly supported
values are . (dot), + (plus sign), x (multiplication sign), *
(asterisk), s (square), o (circle), ^ (up triangle), and v (down
triangle) and d (diamond).

The following code demonstrates the use of *pointtype* and
*pointsize*:

```
require "ishi"
Ishi.new do
plot([1, 2, 3, 4], [5, 6, 7, 8], pointtype: 1, pointsize: 2)
plot([1, 2, 3, 4], [4, 5, 6, 7], pointtype: "o", pointsize: 2)
plot([1, 2, 3, 4], [3, 4, 5, 6], pointtype: "s", pointsize: 2)
plot([1, 2, 3, 4], [2, 3, 4, 5], pointtype: "^", pointsize: 2)
plot([1, 2, 3, 4], [1, 2, 3, 4], pointtype: "v", pointsize: 2, linecolor: "#88001100")
end
```

The *format* argument is a short string used to specify color, point
and line, simultaneously.

Color is a letter from the set b (blue), g (green), r (red), c (cyan), m (magenta), y (yellow), k (black) or w (white). Point is a letter from the set . (dot), + (plus sign), x (multiplication sign), * (asterisk), s (square), c (circle), ^ (up triangle), v (down triangle) or d (diamond). Line may be - (solid line), -- (dashed line), ! (dash-dot line) and : (dotted line).

Given these rules, the format string "b" is a solid blue line, "or" is
red circles, "^{k:"} is black triangles connected by dotted black lines,
"g-" is a solid green line and "--" is a dashed line.

*format* may also be a single word color name or hexadecimal color
value, in which case point and line may not be specified.

Returning to the first example, the code uses *format* to customize
the plots:

```
require "ishi"
Ishi.new do
plot([1, 2, 3, 4, 5], [1.0, 1.4, 1.9, 2.4, 2.6], "ko", title: "data")
plot("0.4 * x + 0.7", "b--")
end
```

The *format* "ko" could also be expressed explicitly (and much more
verbosely) with the named arguments `linecolor: "black", pointtype: 7`

,
and the *format* "b--" with `linecolor: "blue", dashtype: 2`

.

### scatter

`scatter`

draws scatter plots. It takes data in several formats:

`scatter(xdata, ydata)`

- x values in*xdata*and corresponding y values in*ydata*`scatter(xdata, ydata, zdata)`

- x values in*xdata*, y values in*ydata*, and z values in*zdata*

Chart dimensionality (2D or 3D) is inferred from the data. By default,
`scatter`

places a . (dot) at each point.

All `scatter`

methods/commands accept the optional named arguments
*title*, *style*, *dashtype* (*dt*), *linecolor* (*lc*), *linewidth*
(*lw*), *pointsize* (*ps*), *pointtype* (*pt*), *linestyle* (*ls*) and
*format*.

The following code demonstrates the use of *scatter*:

```
require "ishi"
Ishi.new do
scatter(xdata, zdata, lc: "#4b03a1")
scatter(ydata, zdata, lc: "#b5367a")
end
```

### imshow

`imshow`

displays data as a pseudocolor, heatmapped image:

`imshow(data)`

-*data*is two-dimensional scalar data, which will be rendered as an image

The following code demonstrates the use of *imshow*:

```
require "ishi"
Ishi.new do
palette(:inferno)
imshow(data)
margin(0, 0, 0, 0)
show_colorbox(false)
show_border(false)
show_xtics(false)
show_ytics(false)
show_key(false)
end
```

### charts

`charts`

changes the number of charts in the figure:

`charts(rows, cols)`

-*rows*and*cols*are the number of rows and columns in the figure

By default a figure has one chart. This call changes the number of charts in the figure. The original chart is preserved and becomes the chart in the first row, first column of the new layout.

When called without a block, `charts`

returns the charts in the
figure.

```
require "ishi"
figure = Ishi.new
charts = figure.charts(2, 2)
charts[0].plot([1, 2, 3, 4])
charts[1].plot([1, 1, 3, 2])
charts[2].plot([1, 1, 1, 1])
charts[3].plot([4, 3, 2, 1])
figure.show
```

When called with a block, `charts`

Yields each chart as the default
receiver of the supplied block. Block arguments are *i* (the i-th
chart in the figure), *row* and *col* (the row and column of the
chart).

```
require "ishi"
figure = Ishi.new
figure.charts(2, 2) do |i, row, col|
plot([1, 2, 3, 4].rotate(i), title: "#{row},#{col}")
end
figure.show
```

### Non-numeric labels on the x-axis

`xtics`

sets non-numeric labels for positions on the x-axis.

The following code demonstrates the use of *xtics*:

```
require "ishi"
Ishi.new do
x = [1, 2, 3]
y = [65, 30, 5]
plot(x, y, title: "Visits", style: :boxes, fs: 0.25)
.boxwidth(0.5)
.ylabel("Visits (%)")
.xlabel("Device")
.xtics({
1.0 => "mobile",
2.0 => "desktop",
3.0 => "tablet"
})
end
```

### MXNet::NDArray

`MXNet::NDArray`

is an indexable, multi-dimensional array that
efficiently supports numerical operations (transposition, matrix
multiplication, etc.).

All appropriate methods/commands accept instances of `MXNet::NDArray`

with the correct shape (1-dimensional vectors for x-values, for
example).

```
require "ishi"
require "mxnet"
m = MXNet::NDArray.array(
[[0, 0, 1, 1],
[1, 2, 3, 4],
[1, 1, 2, 2]]
)
figure = Ishi.new
charts = figure.charts(1, 2)
charts[0].plot(m[0], m[1], m[2])
charts[1].plot(m[.., 0], m[.., 1], m[.., 2])
figure.show
```

See MXNet.cr for more information on the complete library.

### Extensions

By default, Ishi pops open a window to display charts. However, Ishi
comes with extensions that display charts as text, HTML or inline
images in the terminal; or that render charts to other `IO`

destinations. Note: inline image display only works with
ITerm2.

To plot the *sin(x)* function as text:

```
require "ishi/text" # or "ishi/html" or "ishi/iterm2"
Ishi.new do
plot("sin(x)")
end
```

This produces:

```
1 +--------------------------------------------------------------------+
| * * + * ** + * * |
0.8 |-+ * * * * sin(x* *******-|
| * * * * * * |
0.6 |-+ * * * * * * +-|
| * * * * * * |
0.4 |*+ * * * * * *+-|
|* * * * * * * |
0.2 |*+ * * * * * *+-|
| * * * * * * * |
0 |-* * * * * * *-|
| * * * * * * *|
-0.2 |-+* * * * * * +*|
| * * * * * * *|
-0.4 |-+* * * * * * +*|
| * * * * * * |
-0.6 |-+ * * * * * * +-|
| * * * * * * |
-0.8 |-+ * * * * * * +-|
| * * + ** * + * * |
-1 +--------------------------------------------------------------------+
-10 -5 0 5 10
```

## Contributors

- Todd Sundsted - creator and maintainer
- lbarasti - contributor