Spaces:
Paused
Paused
Update app.py
Browse files
app.py
CHANGED
@@ -396,13 +396,13 @@ def fetch_articles_from_page(url):
|
|
396 |
response = requests.get(url)
|
397 |
response.raise_for_status()
|
398 |
soup = BeautifulSoup(response.content, 'html.parser')
|
399 |
-
articles = soup.find_all('div', class_='
|
400 |
return articles, soup
|
401 |
|
402 |
def extract_articles(articles):
|
403 |
article_data = []
|
404 |
for article in articles:
|
405 |
-
title_div = article.find('
|
406 |
title = title_div.get_text(strip=True) if title_div else "No Title"
|
407 |
date_div = article.find('div', class_='entry-date gt-meta')
|
408 |
date = date_div.get_text(strip=True) if date_div else "No Date"
|
@@ -413,7 +413,7 @@ def extract_articles(articles):
|
|
413 |
article_response = requests.get(link)
|
414 |
article_response.raise_for_status()
|
415 |
article_soup = BeautifulSoup(article_response.content, 'html.parser')
|
416 |
-
article_content_div = article_soup.find('div', class_='entry-
|
417 |
article_content = article_content_div.get_text(strip=True) if article_content_div else "No content found"
|
418 |
article_data.append({
|
419 |
'title': title,
|
|
|
396 |
response = requests.get(url)
|
397 |
response.raise_for_status()
|
398 |
soup = BeautifulSoup(response.content, 'html.parser')
|
399 |
+
articles = soup.find_all('div', class_='entry-post gt-box-shadow-2')
|
400 |
return articles, soup
|
401 |
|
402 |
def extract_articles(articles):
|
403 |
article_data = []
|
404 |
for article in articles:
|
405 |
+
title_div = article.find('h2', class_='entry-title')
|
406 |
title = title_div.get_text(strip=True) if title_div else "No Title"
|
407 |
date_div = article.find('div', class_='entry-date gt-meta')
|
408 |
date = date_div.get_text(strip=True) if date_div else "No Date"
|
|
|
413 |
article_response = requests.get(link)
|
414 |
article_response.raise_for_status()
|
415 |
article_soup = BeautifulSoup(article_response.content, 'html.parser')
|
416 |
+
article_content_div = article_soup.find('div', class_='entry-content')
|
417 |
article_content = article_content_div.get_text(strip=True) if article_content_div else "No content found"
|
418 |
article_data.append({
|
419 |
'title': title,
|