tanner/qotnews
1
0
Fork 1
Code Issues Pull Requests Releases Wiki Activity

Remove get first image

Browse Source
This commit is contained in:
Tanner Collin 2020-07-04 00:27:15 +00:00
parent 850b30e353
commit f1c846acd0
1 changed files with 0 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
apiserver/feed.py
View File

@ -56,19 +56,6 @@ def get_article(url):
logging.error('Problem getting article: {}'.format(str(e))) logging.error('Problem getting article: {}'.format(str(e)))
return '' return ''
def get_first_image(text):
soup = BeautifulSoup(text, features='html.parser')
try:
first_img = soup.find('img')
url = first_img['src']
headers = {'User-Agent': 'Twitterbot/1.0'}
length = requests.get(url, headers=headers, timeout=4).headers['content-length']
if int(length) > 1000000: raise
return url
except:
return ''
def get_content_type(url): def get_content_type(url):
try: try:
headers = {'User-Agent': 'Twitterbot/1.0'} headers = {'User-Agent': 'Twitterbot/1.0'}
@ -130,8 +117,6 @@ def update_story(story, is_manual=False):
story['text'] = get_article(story['url']) story['text'] = get_article(story['url'])
if not story['text']: return False if not story['text']: return False
story['img'] = get_first_image(story['text'])
return True return True
if __name__ == '__main__': if __name__ == '__main__':
@ -147,7 +132,5 @@ if __name__ == '__main__':
a = get_article('https://blog.joinmastodon.org/2019/10/mastodon-3.0/') a = get_article('https://blog.joinmastodon.org/2019/10/mastodon-3.0/')
print(a) print(a)
u = get_first_image(a)
print(u)
print('done') print('done')