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

Ignore new Tildes posts and handle deleted ones

Browse Source
This commit is contained in:
Tanner Collin 2019-09-24 08:21:26 +00:00
parent 0a1ebaa8b8
commit 0496fbba45
1 changed files with 11 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
apiserver/feeds/tildes.py
View File

@ -62,14 +62,18 @@ def story(ref):
html = api(API_ITEM(ref)) html = api(API_ITEM(ref))
if not html: return False if not html: return False
if 'Topic deleted by author' in html: return False
soup = BeautifulSoup(html, features='html.parser') soup = BeautifulSoup(html, features='html.parser')
a = soup.find('article', class_='topic-full') a = soup.find('article', class_='topic-full')
h = a.find('header') h = a.find('header')
lu = h.find('a', class_='link-user')
error = a.find('div', class_='text-error')
if error:
if 'deleted' in error.string or 'removed' in error.string:
return False
s = {} s = {}
s['author'] = str(h.find('a', class_='link-user').string) s['author'] = str(lu.string if lu else 'unknown user')
s['author_link'] = SITE_AUTHOR_LINK(s['author']) s['author_link'] = SITE_AUTHOR_LINK(s['author'])
s['score'] = int(h.find('span', class_='topic-voting-votes').string) s['score'] = int(h.find('span', class_='topic-voting-votes').string)
s['date'] = unix(h.find('time')['datetime']) s['date'] = unix(h.find('time')['datetime'])
@ -85,6 +89,9 @@ def story(ref):
ch = a.find('header', class_='topic-comments-header') ch = a.find('header', class_='topic-comments-header')
s['num_comments'] = int(ch.h2.string.split(' ')[0]) if ch else 0 s['num_comments'] = int(ch.h2.string.split(' ')[0]) if ch else 0
if s['score'] < 8 and s['num_comments'] < 6:
return False
td = a.find('div', class_='topic-full-text') td = a.find('div', class_='topic-full-text')
if td: if td:
s['text'] = td.encode_contents().decode() s['text'] = td.encode_contents().decode()
@ -102,7 +109,7 @@ if __name__ == '__main__':
#print(self_post) #print(self_post)
#li_comment = story('gqx') #li_comment = story('gqx')
#print(li_comment) #print(li_comment)
broken = story('h23') broken = story('hmi')
print(broken) print(broken)
# make sure there's no self-reference # make sure there's no self-reference