AWeirdDev commited on
Commit
7a093ca
1 Parent(s): 470c459

release scraper code

Browse files
Files changed (1) hide show
  1. main.py +64 -0
main.py ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import requests
2
+ from datasets import Dataset
3
+ from selectolax.lexbor import LexborHTMLParser
4
+
5
+ # How many pages to seek for article recommendations?
6
+ # (https://www.storm.mg/articles/{page_id})
7
+ N_PAGES_OF_ARTICLES_RECOMMENDATIONS = 100
8
+
9
+ base_url = "https://www.storm.mg/articles/%i"
10
+ user_agent = (
11
+ # use mine, or put your user agent here
12
+ "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) "
13
+ "Chrome/121.0.0.0 Safari/537.36 OPR/107.0.0.0"
14
+ )
15
+
16
+ def read_article(link: str):
17
+ """Read an article on www.storm.mg."""
18
+ r = requests.get(link, headers={ "User-Agent": user_agent })
19
+ r.raise_for_status()
20
+
21
+ contents = []
22
+ parser = LexborHTMLParser(r.text)
23
+
24
+ for paragraph in parser.css("p[aid]"):
25
+ contents.append(paragraph.text(separator=" ", strip=True))
26
+
27
+ return contents
28
+
29
+
30
+ def generate_dataset():
31
+ """Generate the dataset."""
32
+ for page_id in range(N_PAGES_OF_ARTICLES_RECOMMENDATIONS):
33
+ r = requests.get(base_url % (page_id + 1), headers={
34
+ "User-Agent": user_agent
35
+ })
36
+ r.raise_for_status()
37
+
38
+ parser = LexborHTMLParser(r.text)
39
+ articles = parser.css(".category_cards_wrapper .category_card.card_thumbs_left")
40
+
41
+ for article in articles:
42
+ image = article.css_first("img").attributes['src']
43
+ title = article.css_first(".card_title").text()
44
+ tag = article.css_first(".tags_wrapper a").text()
45
+
46
+ info = article.css_first("p.card_info.right")
47
+ author = info.css_first(".info_author").text()
48
+ timestamp = info.css_first(".info_time").text()
49
+ link = article.css_first(".link_title").attributes['href']
50
+
51
+ yield {
52
+ "image": image,
53
+ "title": title,
54
+ "content": "\n".join(read_article(link)),
55
+ "tag": tag,
56
+ "author": author,
57
+ "timestamp": timestamp,
58
+ "link": link
59
+ }
60
+
61
+ dataset = Dataset.from_generator(generate_dataset)
62
+ dataset.save_to_disk(
63
+ f"storm-org-articles-{20 * N_PAGES_OF_ARTICLES_RECOMMENDATIONS}"
64
+ )