-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathalgorithm.py
180 lines (152 loc) · 7.1 KB
/
algorithm.py
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
import os
import pandas as pd
import Stemmer
from sklearn.feature_extraction.text import TfidfVectorizer
# from sklearn.metrics import f1_score
from sklearn.svm import SVC
from time import time
from pymongo import MongoClient
import logging
import requests
import string
logger = logging.getLogger('newapp')
hdlr = logging.FileHandler('algorithm.log')
formatter = logging.Formatter('%(asctime)s %(levelname)s %(message)s')
hdlr.setFormatter(formatter)
logger.addHandler(hdlr)
logger.setLevel(logging.INFO)
english_stemmer = Stemmer.Stemmer('en')
class StemmedTfidfVectorizer(TfidfVectorizer):
def build_analyzer(self):
analyzer = super(TfidfVectorizer, self).build_analyzer()
return lambda doc: english_stemmer.stemWords(analyzer(doc))
def send_users_notification(database, company_name, article_id):
users_collection = database['users']
query = {'notification_type': 'Companies'}
users = users_collection.find(query)
logger.info('Checking if user follows {}'.format(company_name))
for user in users:
if company_name in user.get('companies'):
url = 'https://api.chatfuel.com/bots/591189a0e4b0772d3373542b/' \
'users/{}/' \
'send?chatfuel_token=vnbqX6cpvXUXFcOKr5RHJ7psSpHDRzO1hXBY8dkvn50ZkZyWML3YdtoCnKH7FSjC' \
'&chatfuel_block_id=5ae5a2d9e4b0f617d6a06eee&last%20name={}&article={}'.\
format(user.get('user_id'), user.get('name'), article_id)
logger.info(user.get('name'))
try:
requests.post(url)
except requests.exceptions.RequestException:
pass
def store_to_database(data, coll):
dbcli = MongoClient('127.0.0.1', 8080)
scrader_db = dbcli['scrader']
scrader_db.authenticate('scrader', 'scr@d3r')
existing_articles = scrader_db[coll]
all_articles = data.to_dict('records')
logger.info('checking to store {} artciles to database'.format(len(all_articles)))
articles_stored = 0
for article in all_articles:
new_article = {}
new_article['title'] = article.get('Title')
new_article['image_url'] = article.get('Image')
new_article['subtitle'] = article.get('Date')
new_article['item_url'] = article.get('Article')
new_article['direction'] = article.get('Sentiment')
new_article['company'] = article.get('Company')
new_article['website'] = article.get('Website')
new_article['website_url'] = article.get('Website url')
new_article['checked'] = False
new_article['appended'] = False
exists = existing_articles.find_one(
{"title": article.get('Title')})
if exists is None:
articles_stored += 1
art_id = existing_articles.insert_one(new_article).inserted_id
# if coll == 'articles':
# send_users_notification(
# scrader_db,
# article.get('Company'),
# art_id
# )
logger.info('storing {} to database'.format(articles_stored))
def train_classifier(clf, X_train, y_train):
""" Fits a classifier to the training data. """
# Start the clock, train the classifier, then stop the clock
start = time()
clf.fit(X_train, y_train)
end = time()
logger.info("Trained model in {:.4f} seconds".format(end - start))
return clf
def predict_labels(clf, features):
# Start the clock, make predictions, then stop the clock
start = time()
y_pred = clf.predict(features)
end = time()
logger.info("Made predictions in {:.4f} seconds.".format(end - start))
return y_pred
def run_algorithm(filename):
dbcli = MongoClient('127.0.0.1', 8080)
db = dbcli['scrader']
db.authenticate('scrader', 'scr@d3r')
collection = db['scraper_companies']
scraper_companies = list(collection.find({}, {'_id': False}))
stop_words_list = []
for comp in scraper_companies:
stop_words_list.extend(comp.get('synonims'))
data = pd.read_csv('./scraderdata.csv', sep=',', encoding='utf-8')
for k in range(len(data)):
data["title"][k] = data["title"][k].lower()
for i in range(len(data)):
data["title"][i] = data["title"][i].encode('utf-8').translate(None, string.punctuation)
for j in range(len(data)):
data["title"][j] = ''.join([i for i in data["title"][j] if not i.isdigit()])
data1 = data['title']
data2 = data['direction']
data = pd.concat([data1.reset_index(drop=True), data2], axis=1)
data.columns = ['title', 'direction']
data.groupby(['direction']).count().reset_index()
train_set = data
train_set.groupby(['direction']).count().reset_index()
data = data.sample(frac=1)
# g = TfidfVectorizer(encoding='utf-8', min_df=3, max_df=10000,
# ngram_range=(1, 2000), analyzer=u'char',
# max_features=20000)
# g = StemmedTfidfVectorizer(min_df=5, max_df=1000, ngram_range=(1, 6), stop_words=[
# "Amazon", "Uber", "Netflix", "Google", "Boeing", "IBM", "Aig", "Apple", "Ryanair", "Motorolla",
# "Equifax", "Microsoft", "Spotify", "Exxon", "Wells Fargo", "Toyota", "HSBC", "BP",
# "Volkswagen", "BnP Paribas", "Daimler", "Samsung", "AXA", "Vodafone", "Nestle", "Ford", "Metlife",
# "General Motors", "Intel", "Oracle", "Unilever", "Morgan Stanley", "Barclays", "Christian Dior", "3M",
# "Target", "Nintendo", "Tesla", "Panasonic", "ebay", "Kia", "Renault", "Apache", "Philips", "Monsanto",
# "Accenture", "Toshiba", "Baidu", "SKY", "JPMorgan", "JP-Morgan", "P&G", "VW", "BMW", "Benz", "Mercedes",
# "AT&T","Renault","Alibaba", "Citi","Chevron","Wal-mart","Gazprom","Verizon", "Santander","Siemens","Novartis",
# "Goldman","Metlife","Hyundai", "Disney","Prudencial","Qualcomm","Honeywell","ABB","Astrazeneca","Carrefour","Canon",
# "Canon","Aetna"], analyzer=u'word', max_features=5000)
g = StemmedTfidfVectorizer(min_df=5, max_df=1000, ngram_range=(1, 6), stop_words=stop_words_list,
analyzer=u'word', max_features=5000)
X_train = g.fit_transform(data['title']).toarray()
y_train = data['direction']
clf = SVC(kernel='linear', probability=True)
clf = train_classifier(clf, X_train, y_train)
real_data = pd.read_csv(filename, sep=',',
encoding='utf-8')
titles = g.transform(real_data['Title']).toarray()
results = predict_labels(clf, titles)
probabilities = clf.predict_proba(titles)
list_probabilities = probabilities.tolist()
best_probs = []
for prob_combo in list_probabilities:
best_probs.append(max(prob_combo))
real_data['Sentiment'] = results
real_data['Probability'] = best_probs
live_data = real_data[real_data.Probability >= 0.7]
store_to_database(real_data, 'dev_articles')
store_to_database(live_data, 'articles')
try:
os.remove("./ScraderwithSentiment.csv")
except OSError:
pass
real_data.to_csv("./ScraderwithSentiment.csv", encoding='utf-8')
if __name__ == '__main__':
# run as script
_file = "./Scraderlatestnews.csv"
run_algorithm(_file)