Network_Science_Yelp_Dataset/graph.py at master · FlxPr/Network_Science_Yelp_Dataset · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
import pandas as pd
from settings import file_names, split_dates
import networkx as nx
from itertools import chain
import igraph
import itertools
import os
import random


def make_friends_graph(library: str = 'networkx'):
    """
    :return: social network graph
    """
    def get_friends_pairs(user_id, friends):
        return [(user_id, friend) for friend in friends.split(', ')] if type(friends) == str else None

    df = pd.read_csv(file_names['toronto_users'])

    if library.lower() == 'networkx':
        social_network = nx.Graph()
        social_network.add_nodes_from(df.user_id.unique())
        social_network.add_edges_from(
            chain.from_iterable(df.apply(lambda row: get_friends_pairs(row['user_id'], row['friends']), axis=1).dropna())
        )
        return social_network

    elif library.lower() == 'igraph':
        social_network = igraph.Graph()
        social_network.add_vertices(df.user_id.unique())
        social_network.add_edges(chain.from_iterable(
                df.apply(lambda row: get_friends_pairs(row['user_id'], row['friends']), axis=1).dropna())
        )
        return social_network
    else:
        raise ValueError('Please use either "networkx" or "igraph" as library')


def make_user_business_bipartite_graph(weighted=False, minimum_rating=4, igraph_=False, mode='full', bipartite=True):
    """
    :param weighted: assign rating as user-business edge weight
    :param minimum_rating: minimum rating to create user-business edge
    :param _igraph: Set to true to have graph in igraph library. Default: networkx library.
    :param mode: Choose in['train', 'test', 'validation', 'full']. Build graph accordingly, spliting accrdingly to review dates as defined in settings.py
    :param bipartite: this argument is just an helper for the function make_frienships_and_reviews_graph.
    :return: user-business interaction graph
    """
    if minimum_rating > 5 and weighted:
        raise ValueError('Minimum rating must be less than 6')

    df = pd.read_csv(file_names['toronto_reviews_without_text'])
    df = df[df.rating >= minimum_rating]

    if bipartite:
        review_network = nx.Graph()
        review_network.add_nodes_from(df.user_id.unique(), bipartite=0)
        review_network.add_nodes_from(df.business_id.unique(), bipartite=1)
    else:
        review_network = nx.Graph()
        review_network.add_nodes_from(df.user_id.unique())
        review_network.add_nodes_from(df.business_id.unique())

    if weighted:
        if mode.lower() == 'train':
            review_network.add_weighted_edges_from([(user, business, rating) for user, business, rating, date
                                   in zip(df.user_id, df.business_id, df.rating, df.date) if date < split_dates['train']['end']])
        elif mode.lower() == 'test':
            review_network.add_weighted_edges_from([(user, business, rating) for user, business, rating, date
                                   in zip(df.user_id, df.business_id, df.rating, df.date) if ((date > split_dates['test']['begin']) & (date < split_dates['test']['end']))])
        elif mode.lower() == 'validation':
            review_network.add_weighted_edges_from([(user, business, rating) for user, business, rating, date
                                   in zip(
                                       df.user_id, df.business_id, df.rating, df.date) if ((date > split_dates['validation']['begin']) & (date < split_dates['validation']['end']))])
        elif mode.lower() == 'full':
            review_network.add_weighted_edges_from([(user, business, rating) for user, business, rating
                                   in zip(df.user_id, df.business_id, df.rating)])
    else:
        if mode.lower() == 'train':
            review_network.add_edges_from([(user, business) for user, business, date
                                   in zip(df.user_id, df.business_id, df.date) if date < split_dates['train']['end']])
        elif mode.lower() == 'test':
            review_network.add_edges_from([(user, business) for user, business, date
                                   in zip(df.user_id, df.business_id, df.date) if ((date > split_dates['test']['begin']) & (date < split_dates['test']['end']))])
        elif mode.lower() == 'validation':
            review_network.add_edges_from([(user, business) for user, business, date
                                   in zip(
                                       df.user_id, df.business_id, df.date) if ((date > split_dates['validation']['begin']) & (date < split_dates['validation']['end']))])
        elif mode.lower() == 'full':
            review_network.add_edges_from([(user, business) for user, business
                                   in zip(df.user_id, df.business_id)])

    if igraph_:

        nx.write_graphml(review_network, 'review_network_temporary.graphml')
        review_network = igraph.read('review_network_temporary.graphml', format='graphml')
        os.remove("review_network_temporary.graphml")

    return review_network


def make_user_business_bipartite_graph_random_split(weighted=False, minimum_rating=4, igraph_=False, mode='full', bipartite=True):
    """
    :param weighted: assign rating as user-business edge weight
    :param minimum_rating: minimum rating to create user-business edge
    :param _igraph: Set to true to have graph in igraph library. Default: networkx library.
    :param mode: Choose in['train', 'test', 'validation', 'full']. Build graph accordingly, spliting accrdingly to review dates as defined in settings.py
    :param bipartite: this argument is just an helper for the function make_frienships_and_reviews_graph.
    :return: user-business interaction graph
    """
    if minimum_rating > 5 and weighted:
        raise ValueError('Minimum rating must be less than 6')

    df = pd.read_csv(file_names['toronto_reviews_without_text'])
    df = df[df.rating >= minimum_rating]

    if bipartite:
        review_network = nx.Graph()
        review_network.add_nodes_from(df.user_id.unique(), bipartite=0)
        review_network.add_nodes_from(df.business_id.unique(), bipartite=1)
    else:
        review_network = nx.Graph()
        review_network.add_nodes_from(df.user_id.unique())
        review_network.add_nodes_from(df.business_id.unique())

    if weighted:

        edges = [(user, business, rating) for user, business, rating, date
                                   in zip(df.user_id, df.business_id, df.rating, df.date)]
        random.seed(1)
        random.shuffle(edges)
        edges_test = edges[:round(len(edges)*0.2)]
        edges_train = edges[round(len(edges)*0.2):]

        if mode.lower() == 'train':
            review_network.add_weighted_edges_from(edges_train)
        elif mode.lower() == 'test':
            review_network.add_weighted_edges_from(edges_test)
        elif mode.lower() == 'full':
            review_network.add_weighted_edges_from(edges)
        else:
            raise ValueError('pass "full", "train" or "test" as mode')

    else:
        edges = [(user, business) for user, business, date in zip(df.user_id, df.business_id, df.date)]

        random.seed(1)
        random.shuffle(edges)
        edges_test = edges[:round(len(edges)*0.2)]
        edges_train = edges[round(len(edges)*0.2):]

        if mode.lower() == 'train':
            review_network.add_edges_from(edges_train)
        elif mode.lower() == 'test':
            review_network.add_edges_from(edges_test)
        elif mode.lower() == 'full':
            review_network.add_edges_from(edges)
        else:
            raise ValueError('pass "full", "train" or "test" as mode')

    if igraph_:

        nx.write_graphml(review_network, 'review_network_temporary.graphml')
        review_network = igraph.read('review_network_temporary.graphml', format='graphml')
        os.remove("review_network_temporary.graphml")

    return review_network


def make_frienships_and_reviews_graph(weight_ratio=1, minimum_rating=0, igraph_=False, mode='full', random_split = False):
    """
    :param weight_ratio: define the ratio of the weights of frienships over reviews. weight_ratio > 1 gives more importance to reviews.
    :param minimum_rating: minimum rating to create user-business edge
    :param igraph: if true, return a graph in igraph library. Otherwise return in networkx library.
    :return: user-user-business interaction graph
    """

    network = make_friends_graph('networkx')
    for e in network.edges():
        network[e[0]][e[1]]['weight'] = 1
    if random_split:
        network_bipartite = make_user_business_bipartite_graph_random_split(weighted=False, minimum_rating=minimum_rating, igraph_=False, mode=mode, bipartite=False)

    else:
        network_bipartite = make_user_business_bipartite_graph(weighted=False, minimum_rating=minimum_rating, igraph_=False, mode=mode, bipartite=False)

    if minimum_rating > 5 :
        raise ValueError('Minimum rating must be less than 6')

    df = pd.read_csv(file_names['toronto_reviews_without_text'])
    df = df[df.rating >= minimum_rating]

    network.add_nodes_from(df.business_id.unique())


    network.add_weighted_edges_from([(user_id, business_id, weight_ratio) for user_id, business_id in list(network_bipartite.edges())])

    #network.add_weighted_edges_from([(user, business, rating) for user, business, rating
    #                        in zip(df.user_id, df.business_id, itertools.repeat(weight_ratio))])

    if igraph_:

        nx.write_graphml(network, 'network_temporary.graphml')
        network = igraph.read('network_temporary.graphml', format='graphml')
        os.remove("network_temporary.graphml")

    return network