on master help sign in
[fix] google news engine change follow-up
commita2c94895c1fbba0b20a07ee13b1185ae741db2de authorAdam Tauber <asciimoo@gmail.com> dateSun 11 Dec 2016, 01:03 (10 years ago) parenta764ebb
2 files changed+58−132
searx/engines/google_news.py+46−34 tests/unit/engines/test_google_news.py+12−98
searx/engines/google_news.py+46−34
@@ -1,41 +1,56 @@
11"""
22 Google (News)
33
4− @website https://www.google.com
5− @provide-api yes (https://developers.google.com/web-search/docs/),
6− deprecated!
4+ @website https://news.google.com
5+ @provide-api no
76
8− @using-api yes
9− @results JSON
10− @stable yes (but deprecated)
7+ @using-api no
8+ @results HTML
9+ @stable no
1110 @parse url, title, content, publishedDate
1211"""
1312
13+from lxml import html
1414from urllib import urlencode
15−from json import loads
16−from dateutil import parser
1715
1816# search-url
1917categories = ['news']
2018paging = True
2119language_support = True
20+safesearch = True
21+time_range_support = True
22+number_of_results = 10
2223
23−# engine dependent config
24−url = 'https://ajax.googleapis.com/'
25−search_url = url + 'ajax/services/search/news?v=2.0&start={offset}&rsz=large&safe=off&filter=off&{query}&hl={lang}'
24+search_url = 'https://www.google.com/search'\
25+ '?{query}'\
26+ '&tbm=nws'\
27+ '&gws_rd=cr'\
28+ '&{search_options}'
29+time_range_attr = "qdr:{range}"
30+time_range_dict = {'day': 'd',
31+ 'week': 'w',
32+ 'month': 'm'}
2633
2734
2835# do search-request
2936def request(query, params):
30− offset = (params['pageno'] - 1) * 8
3137
32− language = 'en-US'
33− if params['language'] != 'all':
34− language = params['language'].replace('_', '-')
38+ search_options = {
39+ 'start': (params['pageno'] - 1) * number_of_results
40+ }
41+
42+ if params['time_range'] in time_range_dict:
43+ search_options['tbs'] = time_range_attr.format(range=time_range_dict[params['time_range']])
44+
45+ if safesearch and params['safesearch']:
46+ search_options['safe'] = 'on'
3547
36− params['url'] = search_url.format(offset=offset,
37− query=urlencode({'q': query}),
38− lang=language)
48+ params['url'] = search_url.format(query=urlencode({'q': query}),
49+ search_options=urlencode(search_options))
50+
51+ if params['language'] != 'all':
52+ language_array = params['language'].lower().split('_')
53+ params['url'] += '&lr=lang_' + language_array[0]
3954
4055 return params
4156
@@ -44,24 +59,21 @@ def request(query, params):
4459def response(resp):
4560 results = []
4661
47− search_res = loads(resp.text)
48−
49− # return empty array if there are no results
50− if not search_res.get('responseData', {}).get('results'):
51− return []
62+ dom = html.fromstring(resp.text)
5263
5364 # parse results
54− for result in search_res['responseData']['results']:
55− # parse publishedDate
56− publishedDate = parser.parse(result['publishedDate'])
57− if 'url' not in result:
58− continue
59−
60− # append result
61− results.append({'url': result['unescapedUrl'],
62− 'title': result['titleNoFormatting'],
63− 'publishedDate': publishedDate,
64− 'content': result['content']})
65+ for result in dom.xpath('//div[@class="g"]|//div[@class="g _cy"]'):
66+ r = {
67+ 'url': result.xpath('.//div[@class="_cnc"]//a/@href')[0],
68+ 'title': ''.join(result.xpath('.//div[@class="_cnc"]//h3//text()')),
69+ 'content': ''.join(result.xpath('.//div[@class="st"]//text()')),
70+ }
71+
72+ img = result.xpath('.//img/@src')[0]
73+ if img and not img.startswith('data'):
74+ r['img_src'] = img
75+
76+ results.append(r)
6577
6678 # return results
6779 return results
tests/unit/engines/test_google_news.py+12−98
@@ -1,3 +1,5 @@
1+# -*- coding: utf-8 -*-
2+
13from collections import defaultdict
24import mock
35from searx.engines import google_news
@@ -11,16 +13,16 @@ class TestGoogleNewsEngine(SearxTestCase):
1113 dicto = defaultdict(dict)
1214 dicto['pageno'] = 1
1315 dicto['language'] = 'fr_FR'
16+ dicto['time_range'] = 'w'
1417 params = google_news.request(query, dicto)
1518 self.assertIn('url', params)
1619 self.assertIn(query, params['url'])
17− self.assertIn('googleapis.com', params['url'])
1820 self.assertIn('fr', params['url'])
1921
2022 dicto['language'] = 'all'
2123 params = google_news.request(query, dicto)
2224 self.assertIn('url', params)
23− self.assertIn('en', params['url'])
25+ self.assertNotIn('fr', params['url'])
2426
2527 def test_response(self):
2628 self.assertRaises(AttributeError, google_news.response, None)
@@ -34,103 +36,15 @@ class TestGoogleNewsEngine(SearxTestCase):
3436 response = mock.Mock(text='{"data": []}')
3537 self.assertEqual(google_news.response(response), [])
3638
37− json = """
38− {
39− "responseData": {
40− "results": [
41− {
42− "GsearchResultClass": "GnewsSearch",
43− "clusterUrl": "http://news.google.com/news/story?ncl=d2d3t1LMDpNIj2MPPhdTT0ycN4sWM&hl=fr&ned=fr",
44− "content": "This is the content",
45− "unescapedUrl": "http://this.is.the.url",
46− "url": "http://this.is.the.url",
47− "title": "This is the title",
48− "titleNoFormatting": "This is the title",
49− "location": "",
50− "publisher": "Jeux Actu",
51− "publishedDate": "Fri, 30 Jan 2015 11:00:25 -0800",
52− "signedRedirectUrl": "http://news.google.com/",
53− "language": "fr",
54− "image": {
55− "url": "http://i.jeuxactus.com/datas/jeux/d/y/dying-light/vu/dying-light-54cc080b568fb.jpg",
56− "tbUrl": "http://t1.gstatic.com/images?q=tbn:ANd9GcSF4yYrs9Ycw23DGiOSAZ-5SEPXYwG3LNs",
57− "originalContextUrl": "http://www.jeuxactu.com/test-dying-light-sur-ps4-97208.htm",
58− "publisher": "Jeux Actu",
59− "tbWidth": 80,
60− "tbHeight": 30
61− },
62− "relatedStories": [
63− {
64− "unescapedUrl": "http://www.jeuxvideo.com/test/415823/dying-light.htm",
65− "url": "http%3A%2F%2Fwww.jeuxvideo.com%2Ftest%2F415823%2Fdying-light.htm",
66− "title": "<b>Test</b> du jeu Dying Light - jeuxvideo.com",
67− "titleNoFormatting": "Test du jeu Dying Light - jeuxvideo.com",
68− "location": "",
69− "publisher": "JeuxVideo.com",
70− "publishedDate": "Fri, 30 Jan 2015 08:52:30 -0800",
71− "signedRedirectUrl": "http://news.google.com/news/url?sa=T&",
72− "language": "fr"
73− }
74− ]
75− }
76− ]
77− },
78− "responseDetails": null,
79− "responseStatus": 200
80− }
81− """
82− response = mock.Mock(text=json)
39+ html = u"""
40+<div class="g">
41+<div class="ts _V6c _Zmc _XO _knc _d7c"><a class="top _vQb _mnc" href="http://this.is.the.url" onmousedown="return rwt(this,'','','','5','AFQjCNGixEtJGC3qTB9pYFLXlRj8XXwdiA','','0ahUKEwiG7O_M5-rQAhWDtRoKHd0RD5QQvIgBCCwwBA','','',event)"><img class="th _lub" id="news-thumbnail-image-52779299683347" src="data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD/2wCEAAkGBwgHBgkIBwgKCgkLDRYPDQwMDRsUFRAWIB0iIiAdHx8kKDQsJCYxJx8fLT0tMTU3Ojo6Iys/RD84QzQ5OjcBCgoKDQwNGg8PGjclHyU3Nzc3Nzc3Nzc3Nzc3Nzc3Nzc3Nzc3Nzc3Nzc3Nzc3Nzc3Nzc3Nzc3Nzc3Nzc3Nzc3N//AABEIAGQAZAMBIgACEQEDEQH/xAAbAAACAwEBAQAAAAAAAAAAAAAFBgADBAIBB//EADsQAAIBAwIEAwUGBAUFAAAAAAECAwAEEQUhBhIxURNBYRQiMnGBFSORobHBB1Ji4UKS0fDxFiQzNYL/xAAYAQEBAQEBAAAAAAAAAAAAAAACAwEEAP/EACERAAMBAAICAgMBAAAAAAAAAAABAhEhMQMSE2EEQVEi/9oADAMBAAIRAxEAPwDSkKr8JFWIzwnMTlSepFdiBgMiuxEdid6B7QbrEr8qE4L4wGNCIT7u/WifE0eIosHBz0oWnugCjQ4BfEipJHbpK/JG0gDP/KO9XT6Lw7BGDHrofIBIJQ4NXalAs8IDJnFLepS21pmOOFXl679FoY66ZTcNwWxtGzaaqwOf8JFcXupSSr4cupSOp6gnGaDQahJESQwU/wBIxXbX0l0jl+VgduYjJrfj/ofc9MluueV1z3xV2jlX1AlDkY61TY3MAk5LuJSnm4G4+nnTFb2ltEweFRvuCO1ebw8uTYNnQ9iDTtDKJYY8DflFJgXLp8xT5FbJHBGx/lFFM9a4MrQFjnFSilvPGiYaItv1zUrTnwHljgLViqwUDyNFEskKlmIDCs08JMWHfl32INZ8hnIscWII4oWPmdqCg9KLccsUtLZM596g6HmVT6Um9WlvGeXk/gWruCObGFz3pb0/R59Y1D2WDYk5lmfpGP3PpW7iK4aIxY3CqWA7t5UxcG3SQWxsZI2VI0MxM6hedj1Oeo9Kc5M6PPagQ3DdjpcpWeHx2U7s+9LuvPCsr+yoETIwAOgpq1DV59RmuJIpbS2hhOG5zvmlLVXE4OCpcnBKdDRW7ptZgMEuCCPrTJwteGYS2z78vvpnt5j/AH3pWIxRbhZ+XV4h5Orj8s/tTtcE5fI6xKWmjVepYfrX0trQrbxLICDyDavm0EvgXEU2M8jhsfI033fHiyxxSfZkwj+EtjP6VzpPR0E8ICQR0qUuT8Y2DSExxtj69alVwOINy3Esye7zKfnXdpZzzMjSfAO1KGm8UX99qNvbm1SNZZApYnoPOimr6rf2crBLmQoTsFIwo8htUHDRk+N0Zv4nQi2itMHALUAiP3SH0oxJrUtyqJcKk4GDiZeb9a0LqkOADY2nz8FaSeLCkxgsWUVrecWQRXwDQW8DzMpGckdNvPr0orxZfpa2if8AYywPIfDkY4YDYHbHbcVgucf9YWd3CqR+MxUgDC5C7bDy2FecRzXQsxJd3lmJN8wCPCvgkZBBOc1Vr2kccaA7PkuBOm6qzD3+XZh1/UflWG+isoGAgMjPzb8x2Jz2r37Sm8AouynqKFyOQxkk8unqa9KYKawolI529CRRLhr/ANxb46AOT/lNBgSTTPwTL7PqMlwpHOkXKuR3IzTvonPY2WqCa6hiPR3VT9TTfx3pzaRb2IskUJICpBPUgUvDXphjLR5HQ8oq6bii9uOUXExmC/CJBzY/GoTWFXOiKV1BZpvEjJYyEkgbb1Kdvt19j9zk9fuk/wBKlL5foz0Zk0vTWa0N/hyVYqiKcZ26/jSvdX81pqET+0PJCX8Nw3l6H5b19H0pccPWpxu2WO/9R3pQ4s0yzi1K4gcFGuW8WN+b3STny8t6xr9lU+MRe2AQVPukZGKq9qTxzCGJcLzbHIqm2uILSCO2uiyvF7pdh1HUNjtQO0uootbLSmQ25kIfwyOYrnyztmvT/oL4C91M3NHIh96GRZB9D+/T61xxNf6XqbBrC2LSbkhVyfypk1Cew1SP7F0exNjbW8LXeovKwMx5VJCs3/0vToT6EUhcHTqOIEfConIyLy+QNX9MWElYOmkYHDKVx5EYrFIrSNk7imriCFrm7kt+TM8R5ie696y6Do41G5naUlLa2TnkxjJ7L9TR3BOdeCzgbg82fntRvR1mtbpGYe5IvUdiP7isrWDSTuIwFIyewA+tbraC7js4J7WVpYz92VUZMTdeUr28wenX1rXyg+rTC7TkHrXonPeiGk6TFeWeNQuWsLot7kkihoXHrj3l+e4oZJbSxzTxY5mgcrJyHmAx57eXrXOVc1Pawt8dq8rPv5VKww+jWUgbhGymTO8QIPod6y8cadHfaQsgws8GWUjrjO4/erXxa8M2cETZCW6bEdNq513UlsbSO6mh9oRXXmh5+UPkk4Jx54NdUrQ08Qj2cQ1ezMDkpexA7k/EBQ+C3ubOVXliICts3l1rXqEr3+qPeWASK7n55vZ7bHhxDGeQf1coJP4dc13Y8SMiPDeRBiRjp19DUaiofBs3NLkKahcwcNaLJaQMtzquspz3jhuYRRE5EYPc9W/4oDDYGyvLO4upLdB4YzFFJmRQq5HMvcj9qtvI7SPXpGguxqFvFysZSvKHON167gdM1a/EUb6fLYz2FvI01yJZrphmVlyCVHYfLue9dTWyRisotfVpLm8D272suB0mURtjtkV3p2qwWF5P7VbmETAK6ZyrD0odrtq+nTiy1PT47Oc4mWWPBflYbA4OMenWsdnd3FusqxOkyMpVkYdR9a5nLXBer2tQyxRWcjznT5CpfDBMbtjoPl6VlN+sCLNbwMZjIysoAGPQj8OtBre5dJeaPKN5ods0aGoW2o2rQ3i/eHYSpsw+dTc4dng/L9J9WjSsK3Og3V/JeQWvhL4cUcr5Zj5j0JGw9aG6Jpst3NDFLcmzNzA8ts2ATIyZ265Xz/vtWbVbOzgQCW4eJ4oT4I8MkSHyUYHmepJAHY0PWGSbT11CFFSGKQwM6EhmYjOTv2OM7darCyeDl/J8r8nk1sIwXDSRK7jDHrnY586lNuh6Pouu6Xb3ZvE06ZV8KaFCFUuvVgD0zsa8oNE9CmpA/YtoRkhrdPptWPidZ5dGuIraETSye6V64jQB3IH82QD8gcb1gvuIDJp9vaW8JVkRYyZN8kDyArBbXx1XXLO2vb2SKI8yyGBSFRSu+MZJzhc/6V1QsYLeo1R3Oh3dnby6RYvDcWcQE1wUVRM3KASQCd/n39aUkhN7rltFJ1muI0OP6mA/ejeqX1jYaxLpOmzmfT4IQgmOMu5PMzbAdwPpS+l37Pq8NzGvM0Myui9yrAgfpXrfIZ6CkmjXWpa9qcPD9s3hQvIyoGGBGrcvmfP96OcI8YaTpelw2mp6MJkR2cThVkEj9dww28umdsVm0HSotb1O6tbK5v8ATp1R2dmcOo94Aqccp8/yoHxBbrpt39lLMswtCweRVwGdt2/ABR9KbWAX2aNX1231DiCXVru3M5kkLmGR/cIxhV+QAH4UHvbqC5lMsFvFa7Y5Yc4P4k0X1vS7XS9E0znwb+6BmkXkX3E8t8ZzuOpxsawXEa6Td3MKrHLJ4JhdiPgYgc2PUbrn51Nz/WPfoxpOcY8QEjbB3q2KUswzjbzWt+t3Ftcw2Hh28cKwRLCzqP8AydMtjA32rnUtG9jiS4srhJ7OT4JUYZ+TDqDQpCVFpmW/g9llYArvGxHwmrNX0v2HQdKkjcxx3HOZk5jhnGMk9+woRFKUO/Uedbprye8tGiZs28Eikcx+FmBG3ocfkKKWcG1zyPv8PLi2Xh3wpbOSQxTuodFHvA4PbuTUpG07XtW0qA29lMYoy3MVx5kD07AVKxzyZphmvZnWViR/KMeQrvQ4lutQm8bJ8G0lmUZ/xKuRn61KldDACoHYzu7MSx6sepo9pFtE/vMMloyT9cj8sV7UoPsS6G7+HX3Wm6jdLvM0ioWPYLn9SaQFka91NHuTzme4Bkz58zb/AK17Uq1dImg5rcrXPGcUU2GSIxqq+QGA36ml+bMkaXEjM0kw8R89yd6lSpiCfCcEV7fXdpcxq8TWUjjujAggr2NXcJ3k08F9pcpDWslpJNyEfA6gEFe29SpQZqAgbmAY9TvXAY5YZOCRUqV79ifRthucRhXgglxsGkTmIHbNSpUrAn//2Q==" alt="A(z) south témájának képe a következőből: CBC.ca" data-deferred="1" onload="google.aft&amp;&amp;google.aft(this)"></a><div class="_cnc"><h3 class="r _U6c"><a class="l _HId" href="http://this.is.the.url" onmousedown="return rwt(this,'','','','5','AFQjCNGixEtJGC3qTB9pYFLXlRj8XXwdiA','','0ahUKEwiG7O_M5-rQAhWDtRoKHd0RD5QQqQIILSgAMAQ','','',event)">Meet Thuli Madonsela — <em>South</em> Africa's conscience</a></h3><div class="slp"><span class="_tQb _IId">CBC.ca</span><span class="_v5">-</span><span class="f nsa _uQb">9 órával ezelőtt</span></div><div class="st"><em>South</em> African Public Protector</div></div><div class="_Xmc card-section"><a class="_sQb" href="http://www.news24.com/Columnists/Mpumelelo_Mkhabela/who-really-governs-south-africa-20161209" onmousedown="return rwt(this,'','','','5','AFQjCNHhc2MnYSZ5T4COqInzvgoju5k5bA','','0ahUKEwiG7O_M5-rQAhWDtRoKHd0RD5QQuogBCC4oATAE','','',event)">Who really governs <em>South</em> Africa?</a><br><span class="_Wmc _GId">Vélemény</span><span class="_v5">-</span><span class="_tQb _IId">News24</span><span class="_v5">-</span><span class="f nsa _uQb">2016. dec. 8.</span></div><div class="_Vmc"></div></div>
42+</div>
43+ """ # noqa
44+ response = mock.Mock(text=html)
8345 results = google_news.response(response)
8446 self.assertEqual(type(results), list)
8547 self.assertEqual(len(results), 1)
86− self.assertEqual(results[0]['title'], 'This is the title')
48+ self.assertEqual(results[0]['title'], u'Meet Thuli Madonsela \u2014 South Africa\'s conscience')
8749 self.assertEqual(results[0]['url'], 'http://this.is.the.url')
88− self.assertEqual(results[0]['content'], 'This is the content')
89−
90− json = """
91− {
92− "responseData": {
93− "results": [
94− {
95− "GsearchResultClass": "GnewsSearch",
96− "clusterUrl": "http://news.google.com/news/story?ncl=d2d3t1LMDpNIj2MPPhdTT0ycN4sWM&hl=fr&ned=fr",
97− "content": "This is the content",
98− "unescapedUrl": "http://this.is.the.url",
99− "title": "This is the title",
100− "titleNoFormatting": "This is the title",
101− "location": "",
102− "publisher": "Jeux Actu",
103− "publishedDate": "Fri, 30 Jan 2015 11:00:25 -0800",
104− "signedRedirectUrl": "http://news.google.com/news/",
105− "language": "fr",
106− "image": {
107− "url": "http://i.jeuxactus.com/datas/jeux/d/y/dying-light/vu/dying-light-54cc080b568fb.jpg",
108− "tbUrl": "http://t1.gstatic.com/images?q=tbn:b_6f-OSAZ-5SEPXYwG3LNs",
109− "originalContextUrl": "http://www.jeuxactu.com/test-dying-light-sur-ps4-97208.htm",
110− "publisher": "Jeux Actu",
111− "tbWidth": 80,
112− "tbHeight": 30
113− }
114− }
115− ]
116− },
117− "responseDetails": null,
118− "responseStatus": 200
119− }
120− """
121− response = mock.Mock(text=json)
122− results = google_news.response(response)
123− self.assertEqual(type(results), list)
124− self.assertEqual(len(results), 0)
125−
126− json = """
127− {
128− "responseData": {},
129− "responseDetails": null,
130− "responseStatus": 200
131− }
132− """
133− response = mock.Mock(text=json)
134− results = google_news.response(response)
135− self.assertEqual(type(results), list)
136− self.assertEqual(len(results), 0)
50+ self.assertEqual(results[0]['content'], 'South African Public Protector')