Skip to content

Latest commit

 

History

History
104 lines (78 loc) · 4.14 KB

plotting_cli.md

File metadata and controls

104 lines (78 loc) · 4.14 KB
Raw

Introduction

Purpose of this Document: This document describes the detailed design of the new CLI for python control of plotting in the MantidPlot application.

Objectives

The Scientific Steering meeting of 2014 highlighted some major issues around the current plotting in MantidPlot. Some of these issues relate to the command line interface (CLI). A full listing of the points can be found here: http://www.mantidproject.org/SSC_2014_Mantid_General. The objective will be to provide a better CLI following the advice and comments provided by instrument scientists.

Detailed Objectives

  • Want simpler control over the plot options
  • Want a limited number of plot options. For the default new CLI configuration.
  • Want a 'familiar' feeling CLI for plotting. For example MatPlot, or MatPlotLib
  • Want a cohesive command set. Individual tools should not have their own unique style and rule set.
  • We should maintain backwards compatibility with current qtiplot as many interfaces will break if not. At least during the inital roll-out
  • Additional top-level plot command that will inspect the data and plot it in the most sensible form.
  • Plot should take an optional tool input, to select which plotting tool to use.
  • plot{...} variants to have a common set of controls, such as for setting log scale, and controlling individual graphs
  • Overwrite plot{...} variants to return a user-friendly plot object
  • Add plotInstrument option
  • Consider python control for future interface exposure such as VSI and tile view. Although the immediate concerns will not involve any additional exposure of UIs.
  • Want the ability to switch plotting style/implementation. Ensure that no changes are made that would impede us from support more than one plotting interface. For example Horace style plotting.

Plotting Functionality to Replicate

The new plotting CLI must deliver analogues for the following functionality:

Plot Command
plotSpectrum
plotBins
plot
plot2D
plot3D
plotSlice
plotMD
instrumentWindow
mergePlots
waterFallPlot
convertToWaterfall
stemPlot
imageView*
VSI*

*Are actually not yet exposed to python yet.

Solution Overview

  • We will base the new style CLI on MatPlotLib.
  • Support both a MatPlotLib style functional methodology, as well as an OO methodology. MatPlotLib already has a very good model for doing this.
  • We will have to deviate from the MatPlotLib style plotting when it comes to plotting of workspaces, which natively contain both x and y data to show, and are a fundamentally differnet data structure than MatPlotLib understands. Under such cases, we will do our best to preserve the options syntax so that it still fits with MatPlotLib.

Typical Current Example Usage

plot_handle = plotSpectrum(source=[{Workspaces}], indices=[{Indexes}]) 
graph = plot_handle.activeLayer()
graph.logLogAxes()

Prototype example usage

Plot

Plot will choose the most appropriate registered plotting tool to perform the job, unless the tool argument is provided. **kwargs hide the large number of options that would have to be registered for all the plotting tools.

plot(source=[{Workspace}], tool={ToolName}, **kwargs)

Return: return type will determine of the exact tool chosen. For example plot_spectrum and plot_bin (see below) would return a tuple of Line2D objects.

Plot Spectrum

line_2ds = plot_spectrum(source=[{Workspaces}], indices=[{Indexes}])

Return: Matplotlib.pyplot plot commands return a tuple of objects to represent every line plotted. Our plotting tools should do the same.

Plot Instrument

instrument_view = plot_instrument({Workspace}, **kwargs)
render_tab = instrument_view.getInstrumentTab()

Plot Slice

slice_viewer = plot_slice([{Workspace}], **kwargs)