monsoon-nlp commited on
Commit
cb58ca6
1 Parent(s): 603f6e2

add script to scrape from wikinews

Browse files
Files changed (1) hide show
  1. scrape.py +42 -0
scrape.py ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # pip install openai lxml cssselector requests xmltodict
2
+
3
+ from datetime import datetime
4
+ import json
5
+ import lxml.html
6
+ from lxml.cssselect import CSSSelector
7
+
8
+ import requests
9
+ import xmltodict
10
+
11
+ # load from WikiNews English
12
+ r = requests.get("https://en.wikinews.org/w/index.php?title=Special:NewsFeed&feed=atom&categories=Published&count=5")
13
+ data = xmltodict.parse(r.content)
14
+
15
+ outputs = []
16
+ entries = data["feed"]["entry"]
17
+ for en in entries:
18
+ # en["summary"]["#text"]
19
+ # en["title"]
20
+ dtme = datetime.strptime(en["updated"], "%Y-%m-%dT%H:%M:%SZ")
21
+ dt = dtme.strftime("%Y/%m/%d")
22
+
23
+ summ = lxml.html.fromstring(en["summary"]["#text"])
24
+
25
+ selAnchor = CSSSelector('a[rel="nofollow"]')
26
+ foundElements = selAnchor(summ)
27
+ for el in foundElements:
28
+ print(el.get("href"))
29
+
30
+ plaintxt = summ.text_content()
31
+ if 'Have an opinion on this story?' in plaintxt:
32
+ plaintxt = plaintxt[:plaintxt.find('Have an opinion on this story?')]
33
+ # print(plaintxt)
34
+
35
+ outputs.append({ "question_date": dt, "question_url": en["link"]["@href"], })
36
+
37
+ tstamp = datetime.now().strftime("%Y%m%d")
38
+ with open(f"./{tstamp}_qa_public.jsonl", "w") as fi:
39
+ for idx, op in enumerate(outputs):
40
+ op["question_id"] = f"{tstamp}_{idx}"
41
+ op["question_source"] = "WikiNews"
42
+ fi.write(json.dumps(op) + "\n")