File size: 10,402 Bytes
fd33638
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
import os
from datetime import datetime, timedelta
from typing import Dict, List
import pandas as pd
import tweepy
import praw
import googleapiclient.discovery
import pytumblr
from gnews import GNews
import requests
from bs4 import BeautifulSoup
import time
import math


class DataFetch:
    def __init__(self):
        # Load company list and set date range
        self.end_date = datetime.now()
        self.start_date = self.end_date - timedelta(days=1)

        # Initialize API clients
        self.tumblr_client = pytumblr.TumblrRestClient(
            os.getenv("TUMBLR_CONSUMER_KEY"),
            os.getenv("TUMBLR_CONSUMER_SECRET"),
            os.getenv("TUMBLR_OAUTH_TOKEN"),
            os.getenv("TUMBLR_OAUTH_SECRET")
        )
        
        twitter_auth = tweepy.OAuthHandler(os.getenv("TWITTER_API_KEY"), os.getenv("TWITTER_API_SECRET"))
        twitter_auth.set_access_token(os.getenv("TWITTER_ACCESS_TOKEN"), os.getenv("TWITTER_ACCESS_TOKEN_SECRET"))
        self.twitter_api = tweepy.API(twitter_auth)

        self.reddit = praw.Reddit(
            client_id=os.getenv("REDDIT_CLIENT_ID"),
            client_secret=os.getenv("REDDIT_CLIENT_SECRET"),
            user_agent="Sentiment Analysis Bot 1.0"
        )

        self.youtube = googleapiclient.discovery.build("youtube", "v3", developerKey=os.getenv("YOUTUBE_API_KEY"))

    def load_company_list(self, file_path: str) -> List[str]:
        self.company_list = pd.read_csv(file_path)['company_ticker'].tolist()

    def collect_data(self) -> List[Dict]:
        all_data = []
        
        for company in self.company_list:
            print(f"{company}:")
            all_data.extend(self._collect_social_media_data(company))
            all_data.extend(self._collect_news_data(company))
        
        return all_data

    def _collect_social_media_data(self, query: str) -> List[Dict]:
        social_data = []

        print("Collecting Reddit Data")
        social_data.extend(self.collect_reddit_data(query))

        print("Collecting YouTube Data")
        social_data.extend(self.collect_youtube_data(query))

        print("Collecting Tumblr Data")
        social_data.extend(self.collect_tumblr_data(query))

        return social_data

    def _collect_news_data(self, query: str) -> List[Dict]:
        news_data = []

        print("Collecting Google News Data")
        news_data.extend(self.collect_google_news(query))

        print("Collecting Financial Times Data")
        news_data.extend(self.collect_financial_times(query))

        print("Collecting Bloomberg Data")
        news_data.extend(self.collect_bloomberg(query))

        print("Collecting Reuters Data")
        news_data.extend(self.collect_reuters(query))

        print("Collecting WSJ Data")
        # news_data.extend(self.collect_wsj(query))

        print("Collecting Serper Data - StockNews, Yahoo Finance, Insider Monkey, Investor's Business Daily, etc.")
        news_data.extend(self.search_news(query))

        return news_data

    def collect_tumblr_data(self, query: str) -> List[Dict]:
        posts = self.tumblr_client.tagged(query)
        return [{"platform": "Tumblr", "company": query, "page_content": {
            "title": post["blog"]["title"], "content": post["blog"]["description"]}} for post in posts]

    def collect_twitter_data(self, query: str) -> List[Dict]:
        tweets = []
        for tweet in tweepy.Cursor(self.twitter_api.search_tweets, q=query, lang="en",
                                   since=self.start_date, until=self.end_date).items(100):
            tweets.append(tweet._json)
        return [{"platform": "Twitter", "company": query, "page_content": tweet} for tweet in tweets]

    def collect_reddit_data(self, query: str) -> List[Dict]:
        posts = []
        subreddit = self.reddit.subreddit("all")
        for post in subreddit.search(query, sort="new", time_filter="day"):
            post_date = datetime.fromtimestamp(post.created_utc)
            if self.start_date <= post_date <= self.end_date:
                posts.append({"platform": "Reddit", "company": query, "page_content": {
                    "title": post.title, "content": post.selftext}})
        return posts

    def collect_youtube_data(self, query: str) -> List[Dict]:
        request = self.youtube.search().list(
            q=query, type="video", part="id,snippet", maxResults=50,
            publishedAfter=self.start_date.isoformat() + "Z", publishedBefore=self.end_date.isoformat() + "Z"
        )
        response = request.execute()
        return [{"platform": "YouTube", "company": query, "page_content": {
            "title": item["snippet"]["title"], "content": item["snippet"]["description"]}} for item in response['items']]

    def collect_google_news(self, query: str) -> List[Dict]:
        google_news = GNews(language='en', country='US', start_date=self.start_date, end_date=self.end_date)
        articles = google_news.get_news(query)
        return [{"platform": "Google News", "company": query, "page_content": {
            "title": article["title"], "content": article["description"]}} for article in articles]

    def collect_financial_times(self, query: str) -> List[Dict]:
        url = f"https://www.ft.com/search?q={query}&dateTo={self.end_date.strftime('%Y-%m-%d')}&dateFrom={self.start_date.strftime('%Y-%m-%d')}"
        response = requests.get(url)
        soup = BeautifulSoup(response.content, 'html.parser')
        articles = soup.find_all('div', class_='o-teaser__content')
        return [{"platform": "Financial Times", "company": query, "page_content": {
            "title": a.find('div', class_='o-teaser__heading').text.strip(),
            "content": a.find('p', class_='o-teaser__standfirst').text.strip() if a.find('p', class_='o-teaser__standfirst') else ''
        }} for a in articles]

    def collect_bloomberg(self, query: str) -> List[Dict]:
        url = f"https://www.bloomberg.com/search?query={query}"
        response = requests.get(url)
        soup = BeautifulSoup(response.content, 'html.parser')
        articles = soup.find_all('div', class_='storyItem__aaf871c1')
        return [{"platform": "Bloomberg", "company": query, "page_content": {
            "title": a.find('a', class_='headline__3a97424d').text.strip(),
            "content": a.find('p', class_='summary__483358e1').text.strip() if a.find('p', class_='summary__483358e1') else ''
        }} for a in articles]

    def collect_reuters(self, query: str) -> List[Dict]:
        articles = []
        base_url = "https://www.reuters.com/site-search/"
        page = 1
        while True:
            url = f"{base_url}?blob={query}&page={page}"
            response = requests.get(url)
            soup = BeautifulSoup(response.content, 'html.parser')
            results = soup.find_all('li', class_='search-result__item')
            if not results:
                break
            for result in results:
                date_elem = result.find('time', class_='search-result__timestamp')
                if date_elem:
                    date = datetime.strptime(date_elem['datetime'], "%Y-%m-%dT%H:%M:%SZ")
                    if self.start_date <= date <= self.end_date:
                        articles.append({"platform": "Reuters", "company": query, "page_content": {
                            "title": result.find('h3', class_='search-result__headline').text.strip(),
                            "content": result.find('p', class_='search-result__excerpt').text.strip()
                        }})
                    elif date < self.start_date:
                        return articles
            page += 1
            time.sleep(1)
        return articles

    def collect_wsj(self, query: str) -> List[Dict]:
        articles = []
        base_url = "https://www.wsj.com/search"
        page = 1
        while True:
            params = {
                'query': query, 'isToggleOn': 'true', 'operator': 'AND', 'sort': 'date-desc',
                'duration': 'custom', 'startDate': self.start_date.strftime('%Y/%m/%d'),
                'endDate': self.end_date.strftime('%Y/%m/%d'), 'page': page
            }
            response = requests.get(base_url, params=params)
            soup = BeautifulSoup(response.content, 'html.parser')
            results = soup.find_all('article', class_='WSJTheme--story--XB4V2mLz')
            if not results:
                break
            for result in results:
                date_elem = result.find('p', class_='WSJTheme--timestamp--22sfkNDv')
                if date_elem:
                    date = datetime.strptime(date_elem.text.strip(), "%B %d, %Y")
                    if self.start_date <= date <= self.end_date:
                        articles.append({"platform": "Wall Street Journal", "company": query, "page_content": {
                            "title": result.find('h3', class_='WSJTheme--headline--unZqjb45').text.strip(),
                            "content": result.find('p', class_='WSJTheme--summary--lmOXEsbN').text.strip()
                        }})
                    elif date < self.start_date:
                        return articles
            page += 1
            time.sleep(1)
        return articles

    def search_news(self, query: str,cnt=300) -> List[Dict]:
        articles = []
        num_results = cnt

        headers = {
            "X-API-KEY": os.getenv("SERP_API_KEY"),
            "Content-Type": "application/json"
        }
        payload = {"q": f"{query} company news",
            "num": num_results,
            "dateRestrict": 14
            }
        response = requests.post(
            "https://google.serper.dev/news",
            headers=headers,
            json=payload
            )
        # print(response)
        if response.status_code == 200:
            results = response.json().get("news", [])
            for result in results:
                articles.append({"platform": result["source"], "company": query, "page_content": {
                    "title": result["title"],
                    "content": result["snippet"],
                    "link": result["link"]
                }})
        return articles

# Usage Example
if __name__ == "__main__":
    analyzer = DataFetch("company_list.csv")
    data = analyzer.collect_data()
    # Here, data would contain all collected sentiment data for the given companies