[enh] reduce the number of http outgoing connections.

engines that still use http : gigablast, bing image for thumbnails, 1x and dbpedia autocompleter
This commit is contained in:
Alexandre Flament 2015-05-02 11:43:12 +02:00
parent bbd83f5a51
commit 78edc16e66
10 changed files with 25 additions and 10 deletions

View file

@ -22,7 +22,7 @@ paging = True
# search-url
base_url = 'https://www.deviantart.com/'
search_url = base_url+'search?offset={offset}&{query}'
search_url = base_url+'browse/all/?offset={offset}&{query}'
# do search-request
@ -56,6 +56,12 @@ def response(resp):
thumbnail_src = link.xpath('.//img')[0].attrib.get('src')
img_src = regex.sub('/', thumbnail_src)
# http to https, remove domain sharding
thumbnail_src = re.sub(r"https?://(th|fc)\d+.", "https://th01.", thumbnail_src)
thumbnail_src = re.sub(r"http://", "https://", thumbnail_src)
url = re.sub(r"http://(.*)\.deviantart\.com/", "https://\\1.deviantart.com/", url)
# append result
results.append({'url': url,
'title': title,