on master help sign in
Merge pull request #525 from ukwt/master

Add Arch Linux Wiki search engine

commitba16f213b02aaf081ce0a59f86d95196039eb884 authorAdam Tauber <asciimoo@gmail.com> dateThu 24 Mar 2016, 21:36 (11 years ago) parent24ea39d8b7dc2a
4 files changed+252
AUTHORS.rst+1 searx/engines/archlinux.py+141 searx/settings.yml+4 tests/unit/engines/test_archlinux.py+106
AUTHORS.rst+1
@@ -41,3 +41,4 @@ generally made searx better:
4141- @GreenLunar
4242- Noemi Vanyi
4343- Kang-min Liu
44+- Kirill Isakov
searx/engines/archlinux.pynew file+141
@@ -0,0 +1,141 @@
1+# -*- coding: utf-8 -*-
2+
3+"""
4+ Arch Linux Wiki
5+
6+ @website https://wiki.archlinux.org
7+ @provide-api no (Mediawiki provides API, but Arch Wiki blocks access to it
8+ @using-api no
9+ @results HTML
10+ @stable no (HTML can change)
11+ @parse url, title
12+"""
13+
14+from urlparse import urljoin
15+from cgi import escape
16+from urllib import urlencode
17+from lxml import html
18+from searx.engines.xpath import extract_text
19+
20+# engine dependent config
21+categories = ['it']
22+language_support = True
23+paging = True
24+base_url = 'https://wiki.archlinux.org'
25+
26+# xpath queries
27+xpath_results = '//ul[@class="mw-search-results"]/li'
28+xpath_link = './/div[@class="mw-search-result-heading"]/a'
29+
30+
31+# cut 'en' from 'en_US', 'de' from 'de_CH', and so on
32+def locale_to_lang_code(locale):
33+ if locale.find('_') >= 0:
34+ locale = locale.split('_')[0]
35+ return locale
36+
37+# wikis for some languages were moved off from the main site, we need to make
38+# requests to correct URLs to be able to get results in those languages
39+lang_urls = {
40+ 'all': {
41+ 'base': 'https://wiki.archlinux.org',
42+ 'search': '/index.php?title=Special:Search&offset={offset}&{query}'
43+ },
44+ 'de': {
45+ 'base': 'https://wiki.archlinux.de',
46+ 'search': '/index.php?title=Spezial:Suche&offset={offset}&{query}'
47+ },
48+ 'fr': {
49+ 'base': 'https://wiki.archlinux.fr',
50+ 'search': '/index.php?title=Spécial:Recherche&offset={offset}&{query}'
51+ },
52+ 'ja': {
53+ 'base': 'https://wiki.archlinuxjp.org',
54+ 'search': '/index.php?title=特別:検索&offset={offset}&{query}'
55+ },
56+ 'ro': {
57+ 'base': 'http://wiki.archlinux.ro',
58+ 'search': '/index.php?title=Special:Căutare&offset={offset}&{query}'
59+ },
60+ 'tr': {
61+ 'base': 'http://archtr.org/wiki',
62+ 'search': '/index.php?title=Özel:Ara&offset={offset}&{query}'
63+ }
64+}
65+
66+
67+# get base & search URLs for selected language
68+def get_lang_urls(language):
69+ if language in lang_urls:
70+ return lang_urls[language]
71+ return lang_urls['all']
72+
73+# Language names to build search requests for
74+# those languages which are hosted on the main site.
75+main_langs = {
76+ 'ar': 'العربية',
77+ 'bg': 'Български',
78+ 'cs': 'Česky',
79+ 'da': 'Dansk',
80+ 'el': 'Ελληνικά',
81+ 'es': 'Español',
82+ 'he': 'עברית',
83+ 'hr': 'Hrvatski',
84+ 'hu': 'Magyar',
85+ 'it': 'Italiano',
86+ 'ko': '한국어',
87+ 'lt': 'Lietuviškai',
88+ 'nl': 'Nederlands',
89+ 'pl': 'Polski',
90+ 'pt': 'Português',
91+ 'ru': 'Русский',
92+ 'sl': 'Slovenský',
93+ 'th': 'ไทย',
94+ 'uk': 'Українська',
95+ 'zh': '简体中文'
96+}
97+
98+
99+# do search-request
100+def request(query, params):
101+ # translate the locale (e.g. 'en_US') to language code ('en')
102+ language = locale_to_lang_code(params['language'])
103+
104+ # if our language is hosted on the main site, we need to add its name
105+ # to the query in order to narrow the results to that language
106+ if language in main_langs:
107+ query += '(' + main_langs[language] + ')'
108+
109+ # prepare the request parameters
110+ query = urlencode({'search': query})
111+ offset = (params['pageno'] - 1) * 20
112+
113+ # get request URLs for our language of choice
114+ urls = get_lang_urls(language)
115+ search_url = urls['base'] + urls['search']
116+
117+ params['url'] = search_url.format(query=query, offset=offset)
118+
119+ return params
120+
121+
122+# get response from search-request
123+def response(resp):
124+ # get the base URL for the language in which request was made
125+ language = locale_to_lang_code(resp.search_params['language'])
126+ base_url = get_lang_urls(language)['base']
127+
128+ results = []
129+
130+ dom = html.fromstring(resp.text)
131+
132+ # parse results
133+ for result in dom.xpath(xpath_results):
134+ link = result.xpath(xpath_link)[0]
135+ href = urljoin(base_url, link.attrib.get('href'))
136+ title = escape(extract_text(link))
137+
138+ results.append({'url': href,
139+ 'title': title})
140+
141+ return results
searx/settings.yml+4
@@ -34,6 +34,10 @@ outgoing: # communication with search engines
3434# - 1.1.1.2
3535
3636engines:
37+ - name : arch linux wiki
38+ engine : archlinux
39+ shortcut : al
40+
3741 - name : wikipedia
3842 engine : mediawiki
3943 shortcut : wp
tests/unit/engines/test_archlinux.pynew file+106
@@ -0,0 +1,106 @@
1+from collections import defaultdict
2+import mock
3+from searx.engines import archlinux
4+from searx.testing import SearxTestCase
5+
6+domains = {
7+ 'all': 'https://wiki.archlinux.org',
8+ 'de': 'https://wiki.archlinux.de',
9+ 'fr': 'https://wiki.archlinux.fr',
10+ 'ja': 'https://wiki.archlinuxjp.org',
11+ 'ro': 'http://wiki.archlinux.ro',
12+ 'tr': 'http://archtr.org/wiki'
13+}
14+
15+
16+class TestArchLinuxEngine(SearxTestCase):
17+
18+ def test_request(self):
19+ query = 'test_query'
20+ dic = defaultdict(dict)
21+ dic['pageno'] = 1
22+ dic['language'] = 'en_US'
23+ params = archlinux.request(query, dic)
24+ self.assertTrue('url' in params)
25+ self.assertTrue(query in params['url'])
26+ self.assertTrue('wiki.archlinux.org' in params['url'])
27+
28+ for lang, domain in domains.iteritems():
29+ dic['language'] = lang
30+ params = archlinux.request(query, dic)
31+ self.assertTrue(domain in params['url'])
32+
33+ def test_response(self):
34+ response = mock.Mock(text='<html></html>',
35+ search_params={'language': 'en_US'})
36+ self.assertEqual(archlinux.response(response), [])
37+
38+ html = """
39+ <ul class="mw-search-results">
40+ <li>
41+ <div class="mw-search-result-heading">
42+ <a href="/index.php/ATI" title="ATI">ATI</a>
43+ </div>
44+ <div class="searchresult">
45+ Lorem ipsum dolor sit amet
46+ </div>
47+ <div class="mw-search-result-data">
48+ 30 KB (4,630 words) - 19:04, 17 March 2016</div>
49+ </li>
50+ <li>
51+ <div class="mw-search-result-heading">
52+ <a href="/index.php/Frequently_asked_questions" title="Frequently asked questions">
53+ Frequently asked questions
54+ </a>
55+ </div>
56+ <div class="searchresult">
57+ CPUs with AMDs instruction set "AMD64"
58+ </div>
59+ <div class="mw-search-result-data">
60+ 17 KB (2,722 words) - 20:13, 21 March 2016
61+ </div>
62+ </li>
63+ <li>
64+ <div class="mw-search-result-heading">
65+ <a href="/index.php/CPU_frequency_scaling" title="CPU frequency scaling">CPU frequency scaling</a>
66+ </div>
67+ <div class="searchresult">
68+ ondemand for AMD and older Intel CPU
69+ </div>
70+ <div class="mw-search-result-data">
71+ 15 KB (2,319 words) - 23:46, 16 March 2016
72+ </div>
73+ </li>
74+ </ul>
75+ """
76+
77+ expected = [
78+ {
79+ 'title': 'ATI',
80+ 'url': 'https://wiki.archlinux.org/index.php/ATI'
81+ },
82+ {
83+ 'title': 'Frequently asked questions',
84+ 'url': 'https://wiki.archlinux.org/index.php/Frequently_asked_questions'
85+ },
86+ {
87+ 'title': 'CPU frequency scaling',
88+ 'url': 'https://wiki.archlinux.org/index.php/CPU_frequency_scaling'
89+ }
90+ ]
91+
92+ response = mock.Mock(text=html)
93+ response.search_params = {
94+ 'language': 'en_US'
95+ }
96+ results = archlinux.response(response)
97+
98+ self.assertEqual(type(results), list)
99+ self.assertEqual(len(results), len(expected))
100+
101+ i = 0
102+ for exp in expected:
103+ res = results[i]
104+ i += 1
105+ for key, value in exp.iteritems():
106+ self.assertEqual(res[key], value)