Skip to content

Latest commit

 

History

History
99 lines (61 loc) · 4.8 KB

EasyPlot_backends.md

File metadata and controls

99 lines (61 loc) · 4.8 KB
Raw

CMDimData.EasyPlot backends

Supported backends

The following is a list of modules implementing the EasyPlot interface, along with their corresponding plotting backends:

  • EasyPlotInspect: (InspectDR.jl)
  • EasyPlotMPL: (PyPlot.jl/Matplotlib)
  • EasyPlotPlots: (Plots.jl)
  • (BROKEN) EasyPlotQwt: (PyCall.jl/guiqwt Python library)
  • EasyPlotGrace: (GracePlot.jl)

Choosing a backend

NOTE: The term "load time" is used loosely below to indicate time to first plot.

  • InspectDR.jl: One of the fastest supported backends with the second shortest load times (second to xmgrace). InspectDR provides good interactivity. Very responsive, even with moderately-sized (~200k points) datasets.
    • Confirmed to handle 2GB datsets with reasonable speed on older desktop running Windows 10 (drag+pan of data area highly discouraged).
  • Grace/xmgrace: Short load times and fast when dealing with small datasets. GUI feels a bit dated and unfamiliar, but one can readily fine tune almost any visual element to generate publication-quality plots.
  • Matplotlib/PyPlot.jl: Longer load times (connects to Python-based matplotlib.pyplot library). Faster than Grace/xmgrace solution when dealing with moderately-sized datasets (~200k points).
  • Qwt/guiqwt: Longer load times (connects to Python-based guiqwt library). Faster than Matplotlib/PyPlot.jl solution when dealing with moderately-sized datasets (~200k points).
    • Though efficient with moderately-sized datasets, guiqwt appears slow when plotting large a number of traces (ex: eye diagram of a long transient dataset split into many individual traces).
  • Plots.jl/(*.jl): Uniform plotting interface supporting multiple backends.

Importing backends

Before importing any backend, CMDimData.EasyPlot must already be imported.

To import a specific interface module for a given plotting backend, as listed in Supported Backends, you must use the @includepkg macro:

#Import base CMDimData/EasyPlot facilities
using CMDimData
using CMDimData.EasyPlot

#Import a specific EasyPlot interface, and its associated backend module:
CMDimData.@includepkg EasyPlotInspect #To render plots with InspectDR

Configuring backends

EasyPlotInspect

By default, EasyPlotInspect renders inline plots with a data display area of x x y pixels??. To specify a different value, set the EASYPLOTINSPECT_RENDERW/H environment variables NOT YET IMPLEMENTED.

The values of EASYPLOTINSPECT_RENDERW/H can therefore be set from ~/.julia/config/startup.jl` with the following:

ENV["EASYPLOTINSPECT_RENDERW"] = "300"
ENV["EASYPLOTINSPECT_RENDERH"] = "200"

EasyPlotMPL

To specify the default Matplotlib/PyPlot backend (i.e. GUI toolkit), add the following to your ~/.julia/config/startup.jl file:

ENV["EASYPLOTMPL_DEFAULTBACKEND"] = "gtk"

See main PyPlot.jl module for supported backends.

EasyPlotPlots

To specify the default Plots.jl plot rendering tool, add the following to your ~/.julia/config/startup.jl file:

ENV["EASYPLOTPLOTS_RENDERINGTOOL"] = "gr"

See Plots.jl Backends documentation page for more information on supported backends.

EasyPlotQwt (BROKEN)

EasyPlotGrace

By default, EasyPlotGrace renders inline plots at 75 dpi. To specify a different value, set the EASYPLOTGRACE_RENDERDPI environment variable:

The value of EASYPLOTGRACE_RENDERDPI can therefore be set from ~/.julia/config/startup.jl with the following:

ENV["EASYPLOTGRACE_RENDERDPI"] = "200"

Additional information: EasyPlotPlots

Supported "sub"-backends

EasyPlotPlots.jl should support all plot rendering tools (backends) supported by Plots.jl itself.

  • Backends without display capabilities: :hdf5
  • Text/terminal-based backends: :unicodeplots
  • Image-only backends: :pgfplotsx
  • GUI-enabled backends: :gr, :inspectdr, :pyplot
  • Python-based backends: :pyplot
  • Browser-enabled? backends: :plotly, :plotlyjs

See Plots.jl Backends documentation page for more information on supported backends & how to install them.

Sample usage

Sample code can be found here.

Known limitations

  • It appears the different backends of EasyPlotPlots do not all operate seamlessly. Unsurprisingly, there appears to be varying levels of difficulty in getting the backends installed & working properly. Consequently, not all backends have been fully tested (if at all).

  • EasyPlotPlots will not auto-install "backend" modules. The EasyPlotPlots/Project.toml file only registers the main Plots.jl module for installation (or rather will do this, if ever it is converted to a proper package).