on master help sign in
Merge pull request #149 from Cqoicebordel/Flickr-engine

Rework Flickr Engine

commitaf416074102d20f3cfd7cd1cfb5cd0e26c8b1f9c authorAdam Tauber <asciimoo@gmail.com> dateMon 22 Dec 2014, 13:33 (12 years ago) parent2e41bfc17fd4ea
3 files changed+163−31
searx/engines/flickr-noapi.py+102 searx/engines/flickr.py+55−29 searx/settings.yml+6−2
searx/engines/flickr-noapi.pynew file+102
@@ -0,0 +1,102 @@
1+#!/usr/bin/env python
2+
3+## Flickr (Images)
4+#
5+# @website https://www.flickr.com
6+# @provide-api yes (https://secure.flickr.com/services/api/flickr.photos.search.html)
7+#
8+# @using-api no
9+# @results HTML
10+# @stable no
11+# @parse url, title, thumbnail, img_src
12+
13+from urllib import urlencode
14+from json import loads
15+from urlparse import urljoin
16+from lxml import html
17+import re
18+
19+categories = ['images']
20+
21+url = 'https://secure.flickr.com/'
22+search_url = url+'search/?{query}&page={page}'
23+photo_url = 'https://www.flickr.com/photos/{userid}/{photoid}'
24+regex = re.compile(r"\"search-photos-models\",\"photos\":(.*}),\"totalItems\":", re.DOTALL)
25+
26+paging = True
27+
28+def build_flickr_url(user_id, photo_id):
29+ return photo_url.format(userid=user_id,photoid=photo_id)
30+
31+
32+def request(query, params):
33+ params['url'] = search_url.format(query=urlencode({'text': query}),
34+ page=params['pageno'])
35+ return params
36+
37+
38+def response(resp):
39+ results = []
40+
41+ matches = regex.search(resp.text)
42+
43+ if matches == None:
44+ return results
45+
46+ match = matches.group(1)
47+ search_results = loads(match)
48+
49+ if not '_data' in search_results:
50+ return []
51+
52+ photos = search_results['_data']
53+
54+ for photo in photos:
55+
56+ # In paged configuration, the first pages' photos are represented by a None object
57+ if photo == None:
58+ continue
59+
60+ # From the biggest to the lowest format
61+ if 'o' in photo['sizes']:
62+ img_src = photo['sizes']['o']['displayUrl']
63+ elif 'k' in photo['sizes']:
64+ img_src = photo['sizes']['k']['displayUrl']
65+ elif 'h' in photo['sizes']:
66+ img_src = photo['sizes']['h']['displayUrl']
67+ elif 'b' in photo['sizes']:
68+ img_src = photo['sizes']['b']['displayUrl']
69+ elif 'c' in photo['sizes']:
70+ img_src = photo['sizes']['c']['displayUrl']
71+ elif 'z' in photo['sizes']:
72+ img_src = photo['sizes']['z']['displayUrl']
73+ elif 'n' in photo['sizes']:
74+ img_src = photo['sizes']['n']['displayUrl']
75+ elif 'm' in photo['sizes']:
76+ img_src = photo['sizes']['m']['displayUrl']
77+ elif 't' in photo['sizes']:
78+ img_src = photo['sizes']['to']['displayUrl']
79+ elif 'q' in photo['sizes']:
80+ img_src = photo['sizes']['q']['displayUrl']
81+ elif 's' in photo['sizes']:
82+ img_src = photo['sizes']['s']['displayUrl']
83+ else:
84+ continue
85+
86+ url = build_flickr_url(photo['owner']['id'], photo['id'])
87+
88+ title = photo['title']
89+
90+ content = '<span class="photo-author">'+ photo['owner']['username'] +'</span><br />'
91+
92+ if 'description' in photo:
93+ content = content + '<span class="description">' + photo['description'] + '</span>'
94+
95+ # append result
96+ results.append({'url': url,
97+ 'title': title,
98+ 'img_src': img_src,
99+ 'content': content,
100+ 'template': 'images.html'})
101+
102+ return results
searx/engines/flickr.py+55−29
@@ -1,54 +1,80 @@
11#!/usr/bin/env python
22
3+## Flickr (Images)
4+#
5+# @website https://www.flickr.com
6+# @provide-api yes (https://secure.flickr.com/services/api/flickr.photos.search.html)
7+#
8+# @using-api yes
9+# @results JSON
10+# @stable yes
11+# @parse url, title, thumbnail, img_src
12+#More info on api-key : https://www.flickr.com/services/apps/create/
13+
314from urllib import urlencode
4−#from json import loads
5−from urlparse import urljoin
6−from lxml import html
7−from time import time
15+from json import loads
816
917categories = ['images']
1018
11−url = 'https://secure.flickr.com/'
12−search_url = url+'search/?{query}&page={page}'
13−results_xpath = '//div[@class="view display-item-tile"]/figure/div'
19+nb_per_page = 15
20+paging = True
21+api_key= None
22+
23+
24+url = 'https://api.flickr.com/services/rest/?method=flickr.photos.search&api_key={api_key}&{text}&sort=relevance&extras=description%2C+owner_name%2C+url_o%2C+url_z&per_page={nb_per_page}&format=json&nojsoncallback=1&page={page}'
25+photo_url = 'https://www.flickr.com/photos/{userid}/{photoid}'
1426
1527paging = True
1628
29+def build_flickr_url(user_id, photo_id):
30+ return photo_url.format(userid=user_id,photoid=photo_id)
31+
1732
1833def request(query, params):
19− params['url'] = search_url.format(query=urlencode({'text': query}),
20− page=params['pageno'])
21− time_string = str(int(time())-3)
22− params['cookies']['BX'] = '3oqjr6d9nmpgl&b=3&s=dh'
23− params['cookies']['xb'] = '421409'
24− params['cookies']['localization'] = 'en-us'
25− params['cookies']['flrbp'] = time_string +\
26− '-3a8cdb85a427a33efda421fbda347b2eaf765a54'
27− params['cookies']['flrbs'] = time_string +\
28− '-ed142ae8765ee62c9ec92a9513665e0ee1ba6776'
29− params['cookies']['flrb'] = '9'
34+ params['url'] = url.format(text=urlencode({'text': query}),
35+ api_key=api_key,
36+ nb_per_page=nb_per_page,
37+ page=params['pageno'])
3038 return params
3139
3240
3341def response(resp):
3442 results = []
35− dom = html.fromstring(resp.text)
36− for result in dom.xpath(results_xpath):
37− img = result.xpath('.//img')
43+
44+ search_results = loads(resp.text)
3845
39− if not img:
40− continue
46+ # return empty array if there are no results
47+ if not 'photos' in search_results:
48+ return []
49+
50+ if not 'photo' in search_results['photos']:
51+ return []
4152
42− img = img[0]
43− img_src = 'https:'+img.attrib.get('src')
53+ photos = search_results['photos']['photo']
4454
45− if not img_src:
55+ # parse results
56+ for photo in photos:
57+ if 'url_o' in photo:
58+ img_src = photo['url_o']
59+ elif 'url_z' in photo:
60+ img_src = photo['url_z']
61+ else:
4662 continue
4763
48− href = urljoin(url, result.xpath('.//a')[0].attrib.get('href'))
49− title = img.attrib.get('alt', '')
50− results.append({'url': href,
64+ url = build_flickr_url(photo['owner'], photo['id'])
65+
66+ title = photo['title']
67+
68+ content = '<span class="photo-author">'+ photo['ownername'] +'</span><br />'
69+
70+ content = content + '<span class="description">' + photo['description']['_content'] + '</span>'
71+
72+ # append result
73+ results.append({'url': url,
5174 'title': title,
5275 'img_src': img_src,
76+ 'content': content,
5377 'template': 'images.html'})
78+
79+ # return results
5480 return results
searx/settings.yml+6−2
@@ -70,10 +70,14 @@ engines:
7070 shortcut : px
7171
7272 - name : flickr
73− engine : flickr
7473 categories : images
7574 shortcut : fl
76− timeout: 3.0
75+# You can use the engine using the official stable API, but you need an API key
76+# See : https://www.flickr.com/services/apps/create/
77+# engine : flickr
78+# api_key: 'apikey' # required!
79+# Or you can use the html non-stable engine, activated by default
80+ engine : flickr-noapi
7781
7882 - name : general-file
7983 engine : generalfile