[subscribestar] strip whitespace from 'content'
This commit is contained in:
@@ -137,7 +137,7 @@ class SubscribestarExtractor(Extractor):
|
|||||||
"author_nick": text.unescape(extr('>', '<')),
|
"author_nick": text.unescape(extr('>', '<')),
|
||||||
"date" : self._parse_datetime(extr(
|
"date" : self._parse_datetime(extr(
|
||||||
'class="post-date">', '</').rpartition(">")[2]),
|
'class="post-date">', '</').rpartition(">")[2]),
|
||||||
"content" : extr('<body>', '</body>')
|
"content" : extr('<body>', '</body>').strip(),
|
||||||
}
|
}
|
||||||
|
|
||||||
def _parse_datetime(self, dt):
|
def _parse_datetime(self, dt):
|
||||||
@@ -194,5 +194,5 @@ class SubscribestarPostExtractor(SubscribestarExtractor):
|
|||||||
"author_nick": text.unescape(extr('alt="', '"')),
|
"author_nick": text.unescape(extr('alt="', '"')),
|
||||||
"date" : self._parse_datetime(extr(
|
"date" : self._parse_datetime(extr(
|
||||||
'<span class="star_link-types">', '<')),
|
'<span class="star_link-types">', '<')),
|
||||||
"content" : extr('<body>', '</body>')
|
"content" : extr('<body>', '</body>').strip(),
|
||||||
}
|
}
|
||||||
|
|||||||
Reference in New Issue
Block a user