def plot_gossip_rate(self):
sequences = analysis.gossip_events(self.events)
for i in xrange(len(sequences)):
# don't want to bucket by address
sequences[i] = [t for t,src,dst in sequences[i]]
legend = self.leaf_projections
scalar = 1./len(self.unique_addresses)
xyvals = [analysis.frequency_count(self,
sequence,
subdivisions=5,
bucket_scalar = scalar)
for sequence in sequences]
ylabel = "Fraction of nodes initiating gossip"
xlabel = "Time (rounds)"
ipython.analysis_plot_2d_multiple(xyvals, xlabel, ylabel, legend)
def plot_state_change_graph_leaves(self):
import matplotlib.pyplot as plt
import numpy as np
import analysis
saved_projection = self.get_projection()
n = len(self.unique_addresses)
xyvals = []
for leaf_projection in self.leaf_projections:
self.set_projection(leaf_projection)
xyvals += [analysis.compute_changes_per_round(self, subdivisions=1, bucket_scalar = 1./n)]
# restore previous projection
self.set_projection(saved_projection)
xlabel = "MiCA Rounds (%s ms)" % self.runtime_info.round_ms
ylabel = "Fraction of nodes changed state"
legend_labels = self.leaf_projections
ipython.analysis_plot_2d_multiple(xyvals, xlabel, ylabel, legend_labels)
def plot_notable_events(self, ne_suffix):
import matplotlib.pyplot as plt
import numpy as np
ne_categories = {} # key -> sequence
for key, timestamp, address in self.notable_events(ne_suffix):
if key not in ne_categories:
ne_categories[key] = []
ne_categories[key].append(timestamp)
notable_sequences = sorted(ne_categories.items())
xlabel = "Round"
ylabel = "Rate (%s events per round)" % ne_suffix
legend_labels = []
xyvals = []
for key, sequence in notable_sequences:
legend_labels += [key]
xy = analysis.frequency_count(self, sequence, subdivisions=5)
xyvals += [xy]
ipython.analysis_plot_2d_multiple(xyvals, xlabel, ylabel, legend_labels)
def plot_2d(self, x, y, xlabel="x", ylabel = "y"):
ipython.analysis_plot_2d_multiple([(x,y)], xlabel, ylabel, )
