Plots

Plotting interface and wrapper for several plotting packages.

This is under development... please add your wishlist for the plotting interface to issue #1.

First, clone the package, and get any plotting packages you need:

Pkg.clone("https://github.com/JuliaPlot/Plots.jl.git")
Pkg.clone("https://github.com/tbreloff/Qwt.jl.git")   # requires pyqt and pyqwt
Pkg.add("Gadfly")  # might also need to Pkg.checkout("Gadfly") and maybe Colors/Compose... I had trouble with it

Now load it in:

using Plots

Do a plot in Qwt, then save a png:

plotter!(:qwt)
plot(1:10)
savepng(ans, Plots.IMG_DIR * "qwt1.png")

which saves:

Do a plot in Gadfly, then save a png:

plotter!(:gadfly)
plot(1:10)
savepng(ans, Plots.IMG_DIR * "gadfly1.png", 6Gadfly.inch, 4Gadfly.inch)

which saves:

Note that you do not need all underlying packages to use this. I use Requires.jl to perform lazy loading of the modules, so there's no initialization until you call plotter!(). This has an added benefit that you can call using Plots and it should return quickly... no more waiting for a plotting package to load when you don't even use it. :)

julia> tic(); using Plots; toc();
elapsed time: 0.356158445 seconds

julia> tic(); using Gadfly; toc();
WARNING: using Gadfly.Plots in module Main conflicts with an existing identifier.
elapsed time: 3.1334697 seconds

plot and plotter! (proposal)

The main plot command. You must call plotter!(:ModuleName) to set the current plotting environment first. Commands are converted into the relevant plotting commands for that package:

  plotter!(:gadfly)
  plot(1:10)    # this calls `y = 1:10; Gadfly.plot(x=1:length(y), y=y)`
  plotter!(:qwt)
  plot(1:10)    # this calls `Qwt.plot(1:10)`

Use plot to create a new plot object, and plot! to add to an existing one:

  plot(args...; kw...)                  # creates a new plot window, and sets it to be the currentPlot
  plot!(args...; kw...)                 # adds to the `currentPlot`
  plot!(plotobj, args...; kw...)        # adds to the plot `plotobj`

Now that you know which plot object you're updating (new, current, or other), I'll leave it off for simplicity. Here are some various args to supply, and the implicit mapping (AVec == AbstractVector and AMat == AbstractMatrix):

  plot(y::AVec; kw...)                       # one line... x = 1:length(y)
  plot(x::AVec, y::AVec; kw...)              # one line (will assert length(x) == length(y))
  plot(y::AMat; kw...)                       # multiple lines (one per column of x), all sharing x = 1:size(y,1)
  plot(x::AVec, y::AMat; kw...)              # multiple lines (one per column of x), all sharing x (will assert length(x) == size(y,1))
  plot(x::AMat, y::AMat; kw...)              # multiple lines (one per column of x/y... will assert size(x) == size(y))
  plot(x::AVec, f::Function; kw...)          # one line, y = f(x)
  plot(x::AMat, f::Function; kw...)          # multiple lines, yᵢⱼ = f(xᵢⱼ)
  plot(x::AVec, fs::AVec{Function}; kw...)   # multiple lines, yᵢⱼ = fⱼ(xᵢ)
  plot(y::AVec{AVec}; kw...)                 # multiple lines, each with x = 1:length(y[i])
  plot(x::AVec, y::AVec{AVec}; kw...)        # multiple lines, will assert length(x) == length(y[i])
  plot(x::AVec{AVec}, y::AVec{AVec}; kw...)  # multiple lines, will assert length(x[i]) == length(y[i])
  plot(n::Integer; kw...)                    # n lines, all empty (for updating plots)

  # TODO: how do we handle NA values in dataframes?
  plot(df::DataFrame; kw...)                 # one line per DataFrame column, labels == names(df)
  plot(df::DataFrame, columns; kw...)        # one line per column, but on a subset of column names

You can swap out plot for subplot. Each line will go into a separate plot. Use the layout keyword:

  y = rand(100,3)
  subplot(y; layout=(2,2), kw...)           # creates 3 lines going into 3 separate plots, laid out on a 2x2 grid (last row is filled with plot #3)
  subplot(y; layout=(1,3), kw...)           # again 3 plots, all in the same row
  subplot(y; layout=[1,[2,3]])              # pass a nested Array to fully specify the layout.  here the first plot will take up the first row, 
                                            # and the others will share the second row

Other potential shorthands:

  hist(args..., kw...)
  scatter(args..., kw...)
  heatmap(args..., kw...)

Some keyword arguments you can set:

  axis            # :left or :right
  color           # can be a string ("red") or a symbol (:red) or a ColorsTypes.jl Colorant (RGB(1,0,0)) or :auto (which lets the package pick)
  label           # string or symbol, applies to that line, may go in a legend
  width           # width of a line
  linetype        # :line, :step, :stepinverted, :sticks, :dots, :none, :heatmap
  linestyle       # :solid, :dash, :dot, :dashdot, :dashdotdot
  marker          # :none, :ellipse, :rect, :diamond, :utriangle, :dtriangle, :cross, :xcross, :star1, :star2, :hexagon
  markercolor     # same choices as `color`
  markersize      # size of the marker
  title           # string or symbol, title of the plot
  xlabel          # string or symbol, label on the bottom (x) axis
  ylabel          # string or symbol, label on the left (y) axis
  yrightlabel     # string or symbol, label on the right (y) axis
  reg             # true or false, add a regression line for each line
  size            # (Int,Int), resize the enclosing window
  pos             # (Int,Int), move the enclosing window to this position
  windowtitle     # string or symbol, set the title of the enclosing windowtitle
  screen          # Integer, move enclosing window to this screen number (for multiscreen desktops)
  show            # true or false, show the plot (in case you don't want the window to pop up right away)

If you don't include a keyword argument, these are the defaults:

  axis = :left
  color = :auto
  label = automatically generated (y1, y2, ...., or y1 (R), y2 (R) for the right axis)
  width = 2
  linetype = :line
  linestype = :solid
  marker = :none
  markercolor = :auto
  markersize = 5
  title = ""
  xlabel = ""
  ylabel = ""
  yrightlabel = ""
  reg = false
  size = (800,600)
  pos = (0,0)
  windowtitle = ""
  screen = 1
  show = true

When plotting multiple lines, you can give every line the same trait by using the singular, or add an "s" to pluralize. (yes I know it's not gramatically correct, but it's easy to use and implement)

  plot(rand(100,2); colors = [:red, RGB(.5,.5,0)], axiss = [:left, :right], width = 5)  # note the width=5 is applied to both lines

TODO

• Plot vectors/matrices
• Plot DataFrames
• Subplots
• Histograms
• 3D plotting
• Scenes/Drawing
• Graphs
• Interactivity (GUIs)
• Gadfly.jl
• PyPlot.jl
• Winston.jl
• Gaston.jl
• GLPlot.jl
• Qwt.jl
• Bokeh.jl
• Plotly.jl
• GoogleCharts.jl
• Vega.jl
• PLplot.jl
• TextPlots.jl
• ASCIIPlots.jl
• Sparklines.jl
• UnicodePlots.jl
• Hinton.jl
• ImageTerm.jl
• GraphViz.jl
• TikzGraphs.jl
• GraphLayout.jl

Author

Thomas Breloff (@tbreloff)