Thesis » QuynhNguyen-MS

import os

import sys

from collections import defaultdict

import numpy as np

import scipy

from scipy import stats

import matplotlib.pyplot as plt

from pdb import set_trace as debugger

class Experiment():

    def __init__(self, num_of_nodes, num_of_snapshots):

        self.scores = np.zeros((num_of_snapshots, num_of_nodes))

        self.index_of_snapshot = 0

        self.num_of_nodes = num_of_nodes

        self.num_of_snapshots = num_of_snapshots

    def add_new_result(self, in_filepath):

        if self.index_of_snapshot == self.num_of_snapshots:

            print "ERROR: the number of snapshots provided is less than the input files"

            sys.exit()

        data = np.loadtxt(in_filepath, delimiter=',', dtype={

                'names': ('node_id', 'score'),

                'formats': ('i4', 'f4')

            })

        for row in data:

            self.update_score(row[0], row[1])

        self.index_of_snapshot += 1

    def update_score(self, node_id, score):

        self.scores[self.index_of_snapshot][node_id] = score

    def summarize(self):

        np.average(self.scores, axis=1)

    def spearman_rank_correlation_coef(self):

        self.time_diff = defaultdict(list)

        for j in range(self.num_of_snapshots):

            for i in range(j, self.num_of_snapshots):

                diff = scipy.stats.spearmanr(self.scores[j], self.scores[i])

                self.time_diff[i-j].append(diff[0])

        max_key = max(self.time_diff.keys()) + 1

        min_diff = [0 for i in range(max_key)]

        max_diff = [0 for i in range(max_key)]

        mean_diff = [0 for i in range(max_key)]

        for i, value in self.time_diff.iteritems():

            min_diff[i] = np.min(value)

            max_diff[i] = np.max(value)

            mean_diff[i] = np.mean(value)

        # Plot

        x_range = sorted(self.time_diff.keys())

        plt.plot(x_range, min_diff, label='min')

        plt.plot(x_range, mean_diff, label='mean')

        plt.plot(x_range, max_diff, label='max')

        plt.ylabel('Spearman rank correlation coefficient')

        plt.xlabel('time diff (?)')

        plt.legend()

        plt.title('FFGraz')

        plt.show()

def all_files_for_network(network_name, dir):

    files = []

    for file in os.listdir(dir):

        prefix = file.split('_')[0]

        if prefix == network_name:

            files.append(os.path.join(dir, file))

    return files

def main():

    dir = 'output'

    network = 'FFGraz'

    files = all_files_for_network(network, dir)

    num_of_snapshots = len(files)

    num_of_nodes = 200

    exp = Experiment(num_of_nodes, num_of_snapshots)

    for file in files:

        exp.add_new_result(file)

    exp.summarize()

    exp.spearman_rank_correlation_coef()

if __name__ == '__main__':

    main()