Thesis » QuynhNguyen-MS

#!/usr/bin/env python

import numpy as np

import matplotlib.pyplot as plt

from pdb import set_trace as debugger

def read_countbcc(input_filepath):

    input_data = np.loadtxt(input_filepath,

        dtype = {

            'names': ('filename', 'topotype', 'num_of_bcc', 'num_of_nodes_in_bcc'),

            'formats': ('S40', 'S6', 'i4', 'S2500'),

        },

        converters = {

            1: lambda s: s.strip(),

            2: lambda s: s.strip(),

            3: lambda s: s.strip(),

        },

        delimiter = ',')

    return input_data

def get_statistical_for_column(col):

    """Returns the min, max, avg, std

    """

    stats = [np.amin(col), np.amax(col), np.average(col), np.std(col)]

    percentage = np.std(col) / np.average(col)

    stats.append(percentage)

    return stats

def average_num_of_bcc(input_data, index_column_name, value_column_name, output_filepath):

    categories = np.unique(input_data[index_column_name])

    result_dict = dict()

    for cat in categories:

        values = input_data[input_data[index_column_name] == cat][value_column_name]

        stats = get_statistical_for_column(values)

        result_dict[cat] = stats

    return result_dict

def plot_num_of_bcc(result_dict, prefixes):

    min_result_dict = dict()

    max_result_dict = dict()

    avg_result_dict = dict()

    for prefix in prefixes:

        min_result_dict[prefix] = [0 for i in range(10)]

        max_result_dict[prefix] = [0 for i in range(10)]

        avg_result_dict[prefix] = [0 for i in range(10)]

    for key, value in result_dict.iteritems():

        topotype = key[0:2]

        num_of_nodes = int(key[2:])

        idx = int(num_of_nodes / 100) - 1

        min_result_dict[topotype][idx] = value[0]

        max_result_dict[topotype][idx] = value[1]

        avg_result_dict[topotype][idx] = value[2]

    # Plot

    x_label = [i for i in range(100, 1001, 100)]

    for prefix in prefixes:

        output_plot_filepath = './output/num_of_bcc_%s.png' % prefix

        y_values = avg_result_dict[prefix]

        plt.plot(x_label, avg_result_dict[prefix], 'ro')

        # plt.axis([, , 0, ])

        plt.xlabel('Number of nodes in a %s graph' % prefix)

        plt.ylabel('Average number of bi-connected component')

        plt.savefig(output_plot_filepath)

        plt.close()

def analyze_nodes_in_bcc(input_data, index_column_name, value_column_name, output_filepath):

    categories = np.unique(input_data[index_column_name])

    for cat in categories:

        rows = input_data[input_data[index_column_name] == cat][value_column_name]

        values = np.array([], dtype='i4')

        for row in rows:

            values = np.append(values, np.asarray(row.split(' '), dtype='i4'))

        output_hist_filepath = './output/histogram_num_of_nodes_%s.png' % cat

        plt.hist(values)

        plt.savefig(output_hist_filepath)

        plt.close()

if __name__ == '__main__':

    input_filepath = 'countbcc_all_graphs.out'

    input_data = read_countbcc(input_filepath)

    avg_bcc_output_filepath = './output/avg_num_of_bcc.out2'

    result_dict = average_num_of_bcc(input_data, 'topotype', 'num_of_bcc', avg_bcc_output_filepath)

    prefixes = ['CN', 'PL']

    plot_num_of_bcc(result_dict, prefixes)

    nodes_in_bcc_output_filepath = './output/nodes_in_bcc.out2'

    analyze_nodes_in_bcc(input_data, 'topotype', 'num_of_nodes_in_bcc', nodes_in_bcc_output_filepath)