diff options
author | sstvinc2 <sstvinc2@gmail.com> | 2017-02-16 16:02:24 -0600 |
---|---|---|
committer | sstvinc2 <sstvinc2@gmail.com> | 2017-02-16 16:02:24 -0600 |
commit | 53e8b692f6374b72238df797bf14e94f0567b331 (patch) | |
tree | aa6bfb1eb1fbe35bd97028d122d30685e591729e /unbiasedFunctions.py | |
parent | 0f2b265ed3c82e674cc714785a9f308be1db9aa1 (diff) |
Added The Guardian to sources
Diffstat (limited to 'unbiasedFunctions.py')
-rw-r--r-- | unbiasedFunctions.py | 17 |
1 files changed, 12 insertions, 5 deletions
diff --git a/unbiasedFunctions.py b/unbiasedFunctions.py index 57c8c6a..de27228 100644 --- a/unbiasedFunctions.py +++ b/unbiasedFunctions.py @@ -22,11 +22,18 @@ def buildArticle(url, sourceName):#, titleDelStart, titleDelEnd, imgDelStart, im f.close()
try:
- #because the quote separator could be ' or ", trim to just before it then lop it off
- img=content.split('og:image" content=')[1][1:].split('>')[0]
- if img[-1]=='/':
- img=img[:-1].strip()
- img=img[:-1]
+ if sourceName=='The Guardian':
+ #The Guardian puts an identifying banner on their og:images
+ #grab the main image from the page instead
+ img=content.split('<img class="maxed', 1)[1]
+ img=img.split('src="', 1)[1].split('"')[0]
+ else:
+ img=content.split('og:image" content=')[1][1:].split('>')[0]
+ if img[-1]=='/':
+ #because the quote separator could be ' or ",
+ #trim to just before it then lop it off
+ img=img[:-1].strip()
+ img=img[:-1]
title=content.split('og:title" content=')[1][1:].split('>')[0]
if title[-1]=='/':
|