NetworkX in pyqtgraph

[JME]

I have been attempting to plot a network, generated from networkX, in a pyqtgraph widget. NetworkX generates graphs for Matplotlib (and by that, I mean...it plots...not well, and not interactive), and I was hoping to be able to use pyqtgraph to display these graphs. Since they are not coordinate based figures, however, I believe that there are some problems passing the network to the plotting functions. I can use networkX to generate [and save] things like .graphML , GEXF, GML,YAML,SparseGraph6, Pajek, and so on, but is there a way to then plot or open these datatypes in a pyqtgraph widget to be able to "zoom around" and look at them?

I thought about using the matplotlib widget within the pyqtgraph plot, but I don't know what this entails, or if it has any added value to just using a matplotlib canvas anyways. 

To make the matplotlib widget, I use the following: 

from PyQt5 import QtWidgets

from matplotlib.figure import Figure

from matplotlib.backends.backend_qt5agg import FigureCanvasQTAgg as Canvas

import matplotlib

# Ensure using PyQt5 backend

matplotlib.use('QT5Agg')

# Matplotlib canvas class to create figure

class MplCanvas(Canvas):

    def __init__(self):

        self.fig = Figure()

        self.ax = self.fig.add_subplot(111)

        self.ax.cla()#test

        Canvas.__init__(self, self.fig)

        Canvas.setSizePolicy(self, QtWidgets.QSizePolicy.Expanding, QtWidgets.QSizePolicy.Expanding)

        Canvas.updateGeometry(self)

# Matplotlib widget

class MplWidget(QtWidgets.QWidget):

    def __init__(self, parent=None):

        QtWidgets.QWidget.__init__(self, parent)   # Inherit from QWidget

        self.canvas = MplCanvas()                  # Create canvas object

        self.vbl = QtWidgets.QVBoxLayout()         # Set box for plotting

        self.vbl.addWidget(self.canvas)

        self.setLayout(self.vbl)

To generate the network in the Matplotlib widget, I do the following: 

        G = nx.Graph()#create a networkx graph

        Comparison_List = pd.read_csv("SampleTracker.csv",index_col=[0]) #just some data filtering for the next few steps

        Comparison_List2 = Comparison_List[Comparison_List["Status"]=='Finalized'].Sample_ID.tolist()

        Comparison_List3 = Comparison_List[Comparison_List["Status"]=='Data Processing Needed'].Sample_ID.tolist()

        fulledgelist=[]

        for i in obj_Network.index.tolist():

            edgelist = eval(obj_Network.loc[i]["Trait_1"])

            if len(edgelist) < len(Comparison_List2):

                edgelist_condensed = edgelist

                for y in edgelist_condensed:

                    Value = (i,y)

                    fulledgelist.append(Value)

        goldlist = [item for item in fulledgelist]

        G.add_nodes_from(obj_Network.index.tolist(),color='yellow') #Add nodes

        G.add_nodes_from(Comparison_List2,color='green')

        G.add_nodes_from(Comparison_List3,color='orange')

        G.add_edges_from(goldlist,weight=0.8) #connect the nodes

        nx.draw_spring(G, with_labels=True,ax=self.widget.canvas.ax)#Altered to display on ax axis

        self.widget.canvas.draw() #this works as an MPL.pyplot object (with correct back-ends enabled)

I believe the problem is the final part, where the "canvas" is a mpl object. Whereas, the pyqtgraph function obviously does not know how to handle the request to "draw". 

Any thoughts or ideas? Thanks for your time.

[Patrick]

Hi,

This won't help with the actual laying out of the graph (as in, I assume networkX can decide the "best" x,y coordinates of the nodes) but pyqtgraph has the GraphItem, which can render a graph network. I'm not sure how you'd go about obtaining the node positions etc from networkX and passing them to the pyqtgraph, but it might be possible to make a bridge between them somehow. Or yeah, get NetworkX to dump out data in some easy format and read it back in and recreate the graph in pyqtgraph.

Have you seen the "Labeled Graph" demo in the examples? (https://github.com/pyqtgraph/pyqtgraph/blob/develop/examples/CustomGraphItem.py) 
Not sure if that will help or not -- apologies I don't have any experience with NetworkX or plotting graphs in pyqtgraph...

With the matplotlib widget, as far as I'm aware it's just a simple wrapper around a matplotlib figure as a QWidget. I don't think it will give you any additional benefits of just plotting in matplotlib.

Patrick