grero / PanGUI

A tool to quickly pan through plots in python

Introduction

PanGUI offers a simple way of panning through a sequence of plots.

Installation

pip install git+https://github.com/grero/DataProcessingTools.git

pip install git+https://github.com/grero/PanGUI.git

conda install --channel grero1980 pangui

Usage

import PanGUI
PanGUI.test()

To create a panning window, use the create_window function, supply a plotobj object accepting two arguments; the first is an index to plot and the second is a matplotlib.figure instance to plot into.

We can see how this works by examining the test function called above. First we generate some random data

class PlotObject(DPObject):
    argsList = ["data", ("title", "test")]

    def __init__(self, *args, **kwargs):
        DPObject.__init__(self, *args ,**kwargs)
        self.data = self.args["data"]
        self.title = self.args["title"]
        self.indexer = self.getindex("trial")
        self.setidx = np.zeros((self.data.shape[0],), dtype=np.int)
        self.current_idx = None

    def plot(self, i=None, getNumEvents=False, getLevels=False, getPlotOpts=False, ax=None, **kwargs):
        """
        This function showcases the structure of a plot function that works with PanGUI.
        """
        # define the plot options that this function understands
        plotopts = {"show": True, "factor": 1.0, "level": "trial","overlay": False,
                    "second_axis": False, "seeds": {"seed1": 1.0, "seed2": 2.0},
                    "color": DPT.objects.ExclusiveOptions(["red","green"], 0)}
        if getPlotOpts:
            return plotopts

        # Extract the recognized plot options from kwargs
        for (k, v) in plotopts.items():
            plotopts[k] = kwargs.get(k, v)

        if getNumEvents:
            # Return the number of events avilable
            if plotopts["level"] == "trial":
                return self.data.shape[0]
            elif plotopts["level"] == "all":
                return 1
        if getLevels:        
            # Return the possible levels for this object
            return ["trial", "all"]
        
        if plotopts["level"] == "all":
            idx = range(self.data.shape[0])
        else:
            idx = i
        if ax is None:
            ax = gca()
        if not plotopts["overlay"]:
            ax.clear()

        if plotopts["show"]:
            f = plotopts["factor"]
            pcolor = plotopts["color"].selected()
            ax.plot(f*self.data[idx, :].T, color=pcolor)
            ax.axvline(plotopts["seeds"]["seed1"])
            ax.axvline(plotopts["seeds"]["seed2"])

            if plotopts["second_axis"]:
                ax2 = ax.twinx()
                ax2.plot(0.5*self.data[i, :].T, color="black")
        return ax

app = PanGUI.create_window(PlotObj())

This will pop-up a plot window that plots the rows of plotobj.data. Clicking prev or next will step through the rows, while a specific row can be selected via the textfield.

Multiple objects can be plotted at the same time, like this:

po1 = PlotObj()
po2 = PlotObj()
po3 = PlotObj()
po4 = PlotObj()
app = PanGUI.create_window([po1, po2, po3, po4], cols=2)

This creates a 2x2 grid of the 4 plot objects.

Axes can be linked using the linkxaxes and linkyaxes keywords.

app = PanGUI.create_window([po1, po2, po3, po4], cols=2,
                           linkxaxes=[0, 0, 2, 2))

This will link the x-axes of the first two objects together, and the x-axes of the second two objects. linkxaxes and linkyaxes should contain the index of the objects to which the axis should be linked. Specifying its own inde, like for the 1st and 3rd object in the above example, is a non-oeration, meaning they specify no link. In other words, the default state, with no axes linking, in the above example is linkxaxes = [0,1,2,3].