Source code for dynamicalab.drawing.plots

import numpy as np

def plot_spectrum(ax, matrix_generator, 
			n_networks=20, 
			nbins=100,
			vals=[], 
			axvline_color="#ef8a62",
			bar_color="#67a9cf", 
			normed=True,
			xlabel=r"$\lambda$",
			ylabel=r"$\rho(\lambda)$",
			label_fontsize=13):
	"""Plot an histogram of the density spectrum of a family of matrices. 

	**Parameters**

	ax : Matplotlib Axes object
		Draw the spectrum in the specified Matplotlib axes.

	matrix_generator : function
		Random generator of the random matrices. Check the example to convert networks to matrices.

	n_networks : int : (default=20)
		Number of networks to sample the statistics
	
	n_bins : int : (defualt=100)
		Number of bins of the histogram.

	vals : list : (default=[])
		Draw a vertical line for each value in the list

	axvline_color : str or list : (default="#ef8a62")
		If a string, each vertical lines will be this color. If a list, it gives the color of each vertical line for elements of ``vals``, and must be the same length as ``vals``.

	bar_color : str : (default="#67a9cf")
		Fill color of the bars of the histogram.
	
	normed : Bool : (default=True)
		If ``True``, the integral of the histogram is equal to one.

	xlabel : str : (default= :math:`\lambda` )
		Label of the X axis.

	ylabel : str :
		Label of the Y axis.

	label_fontsize : int : (default=13)
		Fontsize of the labels
	

	**Example**	

	.. code:: python

		import networkx as nx
		import dynamicalab.drawing as draw
		import matplotlib.pyplot as plt
		import seaborn as sns
		import numpy as np


		def matrix_generator():
			G = nx.erdos_renyi_graph(20,0.5)
			return nx.to_numpy_matrix(G)


		sns.set(style="ticks")
		fig = plt.figure(figsize=(4,2))
		ax = plt.gca()

		draw.plot_spectrum(ax, matrix_generator, 
					n_networks=1000, 
					nbins=80,
					vals=[20*0.5], 
					normed=True,
					label_fontsize=13)

	.. image:: /_static/assets/spectrum_example.png
			:align: center

	
	"""

	# Get stats
	eigenval = np.zeros(1)
	for k in range(0,n_networks):
		A = matrix_generator()
		val, vec = np.linalg.eig(A)
		eigenval = np.concatenate((eigenval, val))

	x = np.real(eigenval)
	y = np.imag(eigenval)
	for i,v in enumerate(vals):
		c = axvline_color
		if isinstance(axvline_color, list):
			c = axvline_color[i]
		ax.axvline(v, color=c, zorder=0)

	ax.hist(x, bins=nbins, color=bar_color, edgecolor="white", linewidth=1, normed=normed)
	
	ax.set_xlabel(xlabel, fontsize=label_fontsize)
	ax.set_ylabel(ylabel, fontsize=label_fontsize)

	return