- GitHub repository: https://github.com/petercorke/ansitable
- Dependencies:
colored
Painless creation of nice-looking tables of data or matrices in Python.
What's new:
0.9.5:
- methods to format table as MarkDown or LaTeX
- work with Python 3.4
0.9.3:
- create matrices as well as tables
- option to suppress color output
Painless creation of nice-looking tables of data for Python.
1 | from ansitable import ANSITable, Column
2 |
3 | table = ANSITable("col1", "column 2 has a big header", "column 3")
4 | table.row("aaaaaaaaa", 2.2, 3)
5 | table.row("bbbbbbbbbbbbb", 5.5, 6)
6 | table.row("ccccccc", 8.8, 9)
7 | table.print()
Line 3 constructs an ANSITable
object and the arguments are a sequence of
column names followed by ANSITable
keyword arguments - there are none in this first example. Since there are three column names this this will be
a 3-column table.
Lines 4-6 add rows, 3 data values for each row.
Line 7 prints the table and yields a tabular display with column widths automatically chosen, and headings and column data all right-justified (default)
col1 column 2 has a big header column 3
aaaaaaaaa 2.2 3
bbbbbbbbbbbbb 5.5 6
ccccccc 8.8 9
By default output is printed to the console (stdout
) but we can also:
- provide a
file
option to.print()
to allow writing to a specified output stream, the default isstdout
. - obtain a multi-line string version of the entire table as
str(table)
.
The more general solution is to provide a sequence of Column
objects which
allows many column specific options to be given, as we shall see later.
For now though, we could rewrite the example above as:
table = ANSITable(
Column("col1"),
Column("column 2 has a big header"),
Column("column 3")
)
or as
table = ANSITable()
table.addcolumn("col1")
table.addcolumn("column 2 has a big header")
table.addcolumn("column 3")
where the keyword arguments to .addcolumn()
are the same as those for
Column
and are given below.
We can specify a Python format()
style format string for any column - by default it
is the general formatting option "{}"
.
You may choose to left or right justify values via the format string, ansitable
provides control over how those resulting strings are justified within the column.
table = ANSITable(
Column("col1"),
Column("column 2 has a big header", "{:.3g}"), # CHANGE
Column("column 3", "{:-10.4f}")
)
table.row("aaaaaaaaa", 2.2, 3)
table.row("bbbbbbbbbbbbb", 5.5, 6)
table.row("ccccccc", 8.8, 9)
table.print()
which yields
col1 column 2 has a big header column 3
aaaaaaaaa 2.2 3.0000
bbbbbbbbbbbbb 5.5 6.0000
ccccccc 8.8 9.0000
Alternatively we can specify the format argument as a function that converts the value to a string.
The data in column 1 is quite long, we might wish to set a maximum column width which
we can do using the width
argument
table = ANSITable(
Column("col1", width=10), # CHANGE
Column("column 2 has a big header", "{:.3g}"),
Column("column 3", "{:-10.4f}")
)
table.row("aaaaaaaaa", 2.2, 3)
table.row("bbbbbbbbbbbbb", 5.5, 6)
table.row("ccccccc", 8.8, 9)
table.print()
which yields
col1 column 2 has a big header column 3
aaaaaaaaa 2.2 3.0000
bbbbbbbbb… 5.5 6.0000
ccccccc 8.8 9.0000
where we see that the data in column 1 has been truncated.
If you don't like the ellipsis you can turn it off, and get to see one more
character, with the ANSITable
option ellipsis=False
. The Unicode ellipsis
character u+2026 is used.
We can add a table border made up of regular ASCII characters
table = ANSITable(
Column("col1"),
Column("column 2 has a big header"),
Column("column 3"),
border="ascii" # CHANGE
)
table.row("aaaaaaaaa", 2.2, 3)
table.row("bbbbbbbbbbbbb", 5.5, 6)
table.row("ccccccc", 8.8, 9)
table.print()
which yields
+--------------+---------------------------+----------+
| col1 | column 2 has a big header | column 3 |
+--------------+---------------------------+----------+
| aaaaaaaaa | 2.2 | 3 |
|bbbbbbbbbbbbb | 5.5 | 6 |
| ccccccc | 8.8 | 9 |
+--------------+---------------------------+----------+
Or we can construct a border using the ANSI box-drawing characters which are supported by most terminal emulators
table = ANSITable(
Column("col1"),
Column("column 2 has a big header"),
Column("column 3"),
border="thick" # CHANGE
)
table.row("aaaaaaaaa", 2.2, 3)
table.row("bbbbbbbbbbbbb", 5.5, 6)
table.row("ccccccc", 8.8, 9)
table.print()
which yields
┏━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ col1 ┃ column 2 has a big header ┃ column 3 ┃
┣━━━━━━━━━━━━━━╋━━━━━━━━━━━━━━━━━━━━━━━━━━━╋━━━━━━━━━━┫
┃ aaaaaaaaa ┃ 2.2 ┃ 3 ┃
┃bbbbbbbbbbbbb ┃ 5.5 ┃ 6 ┃
┃ ccccccc ┃ 8.8 ┃ 9 ┃
┗━━━━━━━━━━━━━━┻━━━━━━━━━━━━━━━━━━━━━━━━━━━┻━━━━━━━━━━┛
Note: this actually looks better on the console than it does in GitHub markdown.
Other border options include "thin", "rounded" (thin with round corners) and "double".
We can change the alignment of data and heading for any column with the alignment flags "<"
(left),
">"
(right) and "^"
(centered).
table = ANSITable(
Column("col1"),
Column("column 2 has a big header", colalign="^"), # CHANGE
Column("column 3"),
border="thick"
)
table.row("aaaaaaaaa", 2.2, 3)
table.row("bbbbbbbbbbbbb", 5.5, 6)
table.row("ccccccc", 8.8, 9)
table.print()
which yields
┏━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃ col1 ┃ column 2 has a big header ┃ column 3 ┃
┣━━━━━━━━━━━━━━╋━━━━━━━━━━━━━━━━━━━━━━━━━━━╋━━━━━━━━━━┫
┃ aaaaaaaaa ┃ 2.2 ┃ 3 ┃
┃bbbbbbbbbbbbb ┃ 5.5 ┃ 6 ┃
┃ ccccccc ┃ 8.8 ┃ 9 ┃
┗━━━━━━━━━━━━━━┻━━━━━━━━━━━━━━━━━━━━━━━━━━━┻━━━━━━━━━━┛
where the data for column 2 has been centered.
Heading and data alignment for any column can be set independently
table = ANSITable(
Column("col1", headalign="<"), # CHANGE
Column("column 2 has a big header", colalign="^"),
Column("column 3", colalign="<"), # CHANGE
border="thick"
)
table.row("aaaaaaaaa", 2.2, 3)
table.row("bbbbbbbbbbbbb", 5.5, 6)
table.row("ccccccc", 8.8, 9)
table.print()
yields
┏━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┓
┃col1 ┃ column 2 has a big header ┃ column 3 ┃
┣━━━━━━━━━━━━━━╋━━━━━━━━━━━━━━━━━━━━━━━━━━━╋━━━━━━━━━━┫
┃ aaaaaaaaa ┃ 2.2 ┃ 3 ┃
┃bbbbbbbbbbbbb ┃ 5.5 ┃ 6 ┃
┃ ccccccc ┃ 8.8 ┃ 9 ┃
┗━━━━━━━━━━━━━━┻━━━━━━━━━━━━━━━━━━━━━━━━━━━┻━━━━━━━━━━┛
where we have left-justified the heading for column 1 and the data for column 3.
If you have the colored
package installed then you can set the foreground and
background color and style (bold, reverse, underlined, dim) of the header and column data, as well as the border color.
table = ANSITable(
Column("col1", headalign="<", colcolor="red", headstyle="underlined"), # CHANGE
Column("column 2 has a big header", colalign="^", colstyle="reverse"), # CHANGE
Column("column 3", colalign="<", colbgcolor="green"), # CHANGE
border="thick", bordercolor="blue" # CHANGE
)
table.row("aaaaaaaaa", 2.2, 3)
table.row("bbbbbbbbbbbbb", 5.5, 6)
table.row("ccccccc", 8.8, 9)
table.print()
which yields
It is also possible to change the color of individual cells in the table
by prefixing the value with a color enclosed in double angle brackets, for example <<red>>
.
table = ANSITable("col1", "column 2 has a big header", "column 3")
table.row("aaaaaaaaa", 2.2, 3)
table.row("<<red>>bbbbbbbbbbbbb", 5.5, 6)
table.row("<<blue>>ccccccc", 8.8, 9)
table.print()
These keyword arguments control the styling of the entire table.
Keyword | Default | Purpose |
---|---|---|
colsep | 2 | Gap between columns (in spaces) |
offset | 0 | Gap at start of each row, shifts the table to the left |
border | no border | Border style: 'ascii', 'thin', 'thick', 'double' |
bordercolor | Border color, see possible values | |
ellipsis | True | Add an ellipsis if a wide column is truncated |
header | True | Include the column header row |
columns | Specify the number of columns if header=False and no header name or Column arguments are given |
|
color | True | Enable color |
- Color is only possible if the
colored
package is installed - If
color
is False then no color escape sequences will be emitted, useful override for tables included in Sphinx documentation.
These keyword arguments control the styling of a single column.
Keyword | Default | Purpose |
---|---|---|
fmt | "{}" |
format string for the column value, or a callable that maps the column value to a string |
width | maximum column width, excess will be truncated | |
colcolor | Text color, see possible values | |
colbgcolor | Text background color, see possible values | |
colstyle | Text style: "bold", "underlined", "reverse", "dim", "blink" | |
colalign | ">" |
Text alignment: ">" (left), "<" (right), "^" (centered) |
headcolor | Heading text color, see possible values | |
headbgcolor | Heading text background color, see possible values | |
headstyle | Heading text style: "bold", "underlined", "reverse", "dim", "blink" | |
headalign | ">" |
Heading text alignment: ">" (left), "<" (right), "^" (centered) |
Note that many terminal emulators do not support the "blink" style.
The main use for this package is to generate tables on the console that are easy to read, but sometimes you might want the table in a different format to include in documentation.
table = ANSITable("col1", "column 2 has a big header", "column 3")
table.row("aaaaaaaaa", 2.2, 3)
table.row("bbbbbbbbbbbbb", -5.5, 6)
table.row("ccccccc", 8.8, -9)
table.print()
can be rendered into Markdown
table.markdown()
| col1 | column 2 has a big header | column 3 |
| ------------: | ------------------------: | -------: |
| aaaaaaaaa | 2.2 | 3 |
| bbbbbbbbbbbbb | -5.5 | 6 |
| ccccccc | 8.8 | -9 |
or LaTex
table.latex()
\begin{tabular}{ |r|r|r| }\hline
\multicolumn{1}{|r|}{col1} & \multicolumn{1}{|r|}{column 2 has a big header} & \multicolumn{1}{|r|}{column 3}\\\hline\hline
aaaaaaaaa & 2.2 & 3 \\
bbbbbbbbbbbbb & -5.5 & 6 \\
ccccccc & 8.8 & -9 \\
\hline
\end{tabular}
In both cases the method returns a string and column alignment is supported. MarkDown doesn't allow the header to have different alignment to the data.
Painless creation of nice-looking matrices for Python.
We can create a formatter for NumPy arrays (1D or 2D)
from ansitable import ANSIMatrix
formatter = ANSIMatrix(style='thick')
and then use it to format a NumPy array
m = np.random.rand(4,4) - 0.5
m[0,0] = 1.23456e-14
formatter.print(m)
yields
┏ ┓
┃ 0 -0.385 -0.106 0.296 ┃
┃ 0.0432 0.339 0.119 -0.468 ┃
┃ 0.405 -0.306 0.0165 -0.439 ┃
┃ 0.203 0.4 -0.499 -0.487 ┃
┗ ┛
we can also add suffixes
formatter.print(m, suffix_super='T', suffix_sub='3')
yields
┏ ┓T
┃ 0 -0.239 0.186 -0.414 ┃
┃ 0.49 0.215 -0.0148 0.0529 ┃
┃ 0.0473 0.0311 0.45 0.394 ┃
┃-0.192 0.193 -0.455 0.0302 ┃
┗ ┛3
By default output is printed to the console (stdout) but we can also:
- provide a
file
option to.print()
to allow writing to a specified output stream, the default isstdout
. - obtain a multi-line string version of the entire table using the
.str()
method instead of.print()
.
The formatter takes additional arguments to control the numeric format and to control the suppression of very small values.
These keyword arguments control the overall styling and operation of the formatter.
Keyword | Default | Purpose |
---|---|---|
style | "thin" |
"thin" , "round" , "thick" , "double" |
fmt | "{:< 10.3g}" |
format for each element |
squish | True | set small elements to zero |
squishtol | 100 | elements less than squishtol * eps are set to zero |
A formatter takes additional arguments to the styling for a particular call.
Keyword | Default | Purpose |
---|---|---|
suffix_super | "" |
superscript suffix text |
suffix_sub | "" |
subscript suffix text |