Creating nice academic-style plot with least typing.
Prerequesties:
- Python >= 3.5 (Necessary)
- Matplotlib == 3.0/3.1 (Necessary)
- Numpy >= 1.13 (Necessary)
- Pandas >= 0.22 (Necessary)
The package is written in pure python and is directly executable. To install it into python library:
pip install -e line
The plotting grammar of Line is similar (but looser) with Gnuplot. The major differences are:
- No quote for filename is needed if column entries are given. i.e.
plot 'filename' 1:2
in Gnuplot is equvalent toplot filename 1:2
in Line; - Supports matlab-style line description (like "r-", "kd--");
- No requirement for frequent
replot
. The replotting is automatically done when there is anything changed (but certainly you can manually replot); - Uniform description of style elements. Line uses CSS model to select and modify element styles, therefore the syntax for getting/setting styles are simple and straightforward;
- A better palette system. Setting palette only requires one command, and one can easily change the color order via modifying
groupid
or usinggroup
command.
The full documentation can be found here, which includes command-line options, command reference, expressions and styles.
Launch in interactive mode:
line
Run a script:
line [scriptname]
Execute commands directly:
line -e [commands]
Plot from file directly:
line -p [filename] (columnx:)(columny)
Line provides various ways to select columns:
line> plot test-data.txt 2 # plot second column against data index
line> plot test-data.txt $0:2 # same as above, since '$0' is data index
line> plot test-data.txt 1:2 # plot second column agianst first column
line> plot test-data.txt 1:2,3 # plot second and third column against first column
line> plot test-data.txt t:y1 # plot column 'y1' against column 't'
line> plot test-data.txt 1:y1 # plot column 'y1' against first column
line> plot 'test-data.txt' # plot remained columns against first column
line> plot 'test-data.txt', 'hist-data.txt' # plot two files
Customize line styles while plotting:
line> plot test-data.txt t:y1 t='second column' # set data title
line> plot test-data.txt t:y1 lw=2 lc=red # plot with linewidth and linecolor set
line> plot test-data.txt 1:2 'rx-' # plot with line descriptor
The data delimiter and existence of head is automatically determined, but can be specified by data-delimiter and data-title options. Append data to existing figure ("hold on"):
line> add test-data.txt 3 # add the third column
Adjust ranges, scales, grids, legends, etc. (the beginning "set" can be omitted):
line> set xrange 0:2 # setting x range, the 'set' can be omitted
line> xrange 0:2 # setting x range
line> xscale log # set x scale
line> ylabel 'y' # set ylabel
line> grid on # show grid
line> grid lt=dash # setting grid style
line> legend off # hide legend
line> legend pos=top,left # set legend position
More setting commands:
line> set tick format='%.3f' # set tick format
line> set label fontfamily=Arial # set label font
line> set line lw=2 # set linewidth of all data lines
line> set line1 lw=2 # only set line0
line> set line:label='data3' color=red # set line with certain label to red
line> set option auto-adjust-range=false # disable automatic range adjust
line> set palette mpl.OrRd # set palette
line> set palette lighter bar # set palette for bars
Its counterparts, "show" command, displays all style parameters.
line> show line1 # show all attributes of line1
line> show line lw # show linewidth of all lines
Load a script:
load [scriptname]
Display figure in a script:
display
Interrupt script running and switch to interactive mode:
input
Note Line will not display the figure unless display
or input
are given. The command input
is useful when one needs to modify scripts.
Line reads .linerc
in user's home directory, which is just a script. You can also use load
command to specify configurations.
More custom style classes can be added to styles/defaults.d.css. Palettes can be added to styles/palettes.json.