webapp.py 10 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326
  1. #!/usr/bin/env python
  2. '''
  3. searx is free software: you can redistribute it and/or modify
  4. it under the terms of the GNU Affero General Public License as published by
  5. the Free Software Foundation, either version 3 of the License, or
  6. (at your option) any later version.
  7. searx is distributed in the hope that it will be useful,
  8. but WITHOUT ANY WARRANTY; without even the implied warranty of
  9. MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  10. GNU Affero General Public License for more details.
  11. You should have received a copy of the GNU Affero General Public License
  12. along with searx. If not, see < http://www.gnu.org/licenses/ >.
  13. (C) 2013- by Adam Tauber, <asciimoo@gmail.com>
  14. '''
  15. import json
  16. import cStringIO
  17. import os
  18. from flask import Flask, request, render_template
  19. from flask import url_for, Response, make_response, redirect
  20. from flask import send_from_directory
  21. from searx import settings
  22. from searx.engines import search, categories, engines, get_engines_stats
  23. from searx.utils import UnicodeWriter
  24. from searx.utils import highlight_content, html_to_text
  25. from flask.ext.babel import Babel
  26. app = Flask(
  27. __name__,
  28. static_folder=os.path.join(os.path.dirname(__file__), 'static'),
  29. template_folder=os.path.join(os.path.dirname(__file__), 'templates')
  30. )
  31. app.secret_key = settings['server']['secret_key']
  32. babel = Babel(app)
  33. #TODO configurable via settings.yml
  34. favicons = ['wikipedia', 'youtube', 'vimeo', 'soundcloud',
  35. 'twitter', 'stackoverflow', 'github']
  36. opensearch_xml = '''<?xml version="1.0" encoding="utf-8"?>
  37. <OpenSearchDescription xmlns="http://a9.com/-/spec/opensearch/1.1/">
  38. <ShortName>searx</ShortName>
  39. <Description>Search searx</Description>
  40. <InputEncoding>UTF-8</InputEncoding>
  41. <LongName>searx meta search engine</LongName>
  42. <Url type="text/html" method="{method}" template="{host}">
  43. <Param name="q" value="{{searchTerms}}" />
  44. </Url>
  45. </OpenSearchDescription>
  46. '''
  47. @babel.localeselector
  48. def get_locale():
  49. locale = request.accept_languages.best_match(settings['locales'].keys())
  50. if request.cookies.get('locale', '') in settings['locales']:
  51. locale = request.cookies.get('locale', '')
  52. if 'locale' in request.args\
  53. and request.args['locale'] in settings['locales']:
  54. locale = request.args['locale']
  55. if 'locale' in request.form\
  56. and request.form['locale'] in settings['locales']:
  57. locale = request.form['locale']
  58. return locale
  59. def get_base_url():
  60. if settings['server']['base_url']:
  61. hostname = settings['server']['base_url']
  62. else:
  63. scheme = 'http'
  64. if request.is_secure:
  65. scheme = 'https'
  66. hostname = url_for('index', _external=True, _scheme=scheme)
  67. return hostname
  68. def render(template_name, **kwargs):
  69. global categories
  70. kwargs['categories'] = ['general']
  71. kwargs['categories'].extend(x for x in
  72. sorted(categories.keys()) if x != 'general')
  73. if not 'selected_categories' in kwargs:
  74. kwargs['selected_categories'] = []
  75. cookie_categories = request.cookies.get('categories', '').split(',')
  76. for ccateg in cookie_categories:
  77. if ccateg in categories:
  78. kwargs['selected_categories'].append(ccateg)
  79. if not len(kwargs['selected_categories']):
  80. kwargs['selected_categories'] = ['general']
  81. return render_template(template_name, **kwargs)
  82. def parse_query(query):
  83. query_engines = []
  84. query_parts = query.split()
  85. if query_parts[0].startswith('-'):
  86. engine_name = query_parts[0][1:].replace('_', ' ')
  87. if engine_name in engines:
  88. query_engines.append({'category': 'none',
  89. 'name': query_parts[0][1:]})
  90. elif query_parts[0].startswith('!'):
  91. category_name = query_parts[0][1:].replace('_', ' ')
  92. if category_name in categories:
  93. query_engines.extend({'category': category_name,
  94. 'name': engine.name}
  95. for engine in categories[category_name])
  96. if len(query_engines):
  97. query = query.replace(query_parts[0], '', 1).strip()
  98. return query, query_engines
  99. @app.route('/', methods=['GET', 'POST'])
  100. def index():
  101. global categories
  102. if request.method == 'POST':
  103. request_data = request.form
  104. else:
  105. request_data = request.args
  106. if not request_data.get('q'):
  107. return render('index.html')
  108. selected_categories = []
  109. query, selected_engines = parse_query(request_data['q'].encode('utf-8'))
  110. if len(selected_engines):
  111. selected_categories = list(set(engine['category']
  112. for engine in selected_engines))
  113. else:
  114. for pd_name, pd in request_data.items():
  115. if pd_name.startswith('category_'):
  116. category = pd_name[9:]
  117. if not category in categories:
  118. continue
  119. selected_categories.append(category)
  120. if not len(selected_categories):
  121. cookie_categories = request.cookies.get('categories', '')
  122. cookie_categories = cookie_categories.split(',')
  123. for ccateg in cookie_categories:
  124. if ccateg in categories:
  125. selected_categories.append(ccateg)
  126. if not len(selected_categories):
  127. selected_categories = ['general']
  128. for categ in selected_categories:
  129. selected_engines.extend({'category': categ,
  130. 'name': x.name}
  131. for x in categories[categ])
  132. results, suggestions = search(query, request, selected_engines)
  133. featured_results = []
  134. for result in results:
  135. if request_data.get('format', 'html') == 'html':
  136. if 'content' in result:
  137. result['content'] = highlight_content(result['content'], query)
  138. result['title'] = highlight_content(result['title'], query)
  139. else:
  140. if 'content' in result:
  141. result['content'] = html_to_text(result['content']).strip()
  142. result['title'] = html_to_text(result['title']).strip()
  143. if len(result['url']) > 74:
  144. url_parts = result['url'][:35], result['url'][-35:]
  145. result['pretty_url'] = '{0}[...]{1}'.format(*url_parts)
  146. else:
  147. result['pretty_url'] = result['url']
  148. for engine in result['engines']:
  149. if engine in favicons:
  150. result['favicon'] = engine
  151. if request_data.get('format') == 'json':
  152. return Response(json.dumps({'query': query, 'results': results}),
  153. mimetype='application/json')
  154. elif request_data.get('format') == 'csv':
  155. csv = UnicodeWriter(cStringIO.StringIO())
  156. keys = ('title', 'url', 'content', 'host', 'engine', 'score')
  157. if len(results):
  158. csv.writerow(keys)
  159. for row in results:
  160. row['host'] = row['parsed_url'].netloc
  161. csv.writerow([row.get(key, '') for key in keys])
  162. csv.stream.seek(0)
  163. response = Response(csv.stream.read(), mimetype='application/csv')
  164. content_disp = 'attachment;Filename=searx_-_{0}.csv'.format(query)
  165. response.headers.add('Content-Disposition', content_disp)
  166. return response
  167. elif request_data.get('format') == 'rss':
  168. response_rss = render(
  169. 'opensearch_response_rss.xml',
  170. results=results,
  171. q=request_data['q'],
  172. number_of_results=len(results),
  173. base_url=get_base_url()
  174. )
  175. return Response(response_rss, mimetype='text/xml')
  176. return render(
  177. 'results.html',
  178. results=results,
  179. q=request_data['q'],
  180. selected_categories=selected_categories,
  181. number_of_results=len(results) + len(featured_results),
  182. featured_results=featured_results,
  183. suggestions=suggestions
  184. )
  185. @app.route('/about', methods=['GET'])
  186. def about():
  187. return render('about.html')
  188. @app.route('/engines', methods=['GET'])
  189. def list_engines():
  190. global categories
  191. return render('engines.html', categs=categories.items())
  192. @app.route('/preferences', methods=['GET', 'POST'])
  193. def preferences():
  194. if request.method == 'POST':
  195. selected_categories = []
  196. locale = None
  197. for pd_name, pd in request.form.items():
  198. if pd_name.startswith('category_'):
  199. category = pd_name[9:]
  200. if not category in categories:
  201. continue
  202. selected_categories.append(category)
  203. elif pd_name == 'locale' and pd in settings['locales']:
  204. locale = pd
  205. resp = make_response(redirect('/'))
  206. if locale:
  207. # cookie max age: 4 weeks
  208. resp.set_cookie(
  209. 'locale', locale,
  210. max_age=60 * 60 * 24 * 7 * 4
  211. )
  212. if selected_categories:
  213. # cookie max age: 4 weeks
  214. resp.set_cookie(
  215. 'categories', ','.join(selected_categories),
  216. max_age=60 * 60 * 24 * 7 * 4
  217. )
  218. return resp
  219. return render('preferences.html',
  220. locales=settings['locales'],
  221. current_locale=get_locale())
  222. @app.route('/stats', methods=['GET'])
  223. def stats():
  224. global categories
  225. stats = get_engines_stats()
  226. return render('stats.html', stats=stats)
  227. @app.route('/robots.txt', methods=['GET'])
  228. def robots():
  229. return Response("""User-agent: *
  230. Allow: /
  231. Allow: /about
  232. Disallow: /stats
  233. Disallow: /engines
  234. """, mimetype='text/plain')
  235. @app.route('/opensearch.xml', methods=['GET'])
  236. def opensearch():
  237. global opensearch_xml
  238. method = 'post'
  239. # chrome/chromium only supports HTTP GET....
  240. if request.headers.get('User-Agent', '').lower().find('webkit') >= 0:
  241. method = 'get'
  242. base_url = get_base_url()
  243. ret = opensearch_xml.format(method=method, host=base_url)
  244. resp = Response(response=ret,
  245. status=200,
  246. mimetype="application/xml")
  247. return resp
  248. @app.route('/favicon.ico')
  249. def favicon():
  250. return send_from_directory(os.path.join(app.root_path, 'static/img'),
  251. 'favicon.png',
  252. mimetype='image/vnd.microsoft.icon')
  253. def run():
  254. from gevent import monkey
  255. monkey.patch_all()
  256. app.run(
  257. debug=settings['server']['debug'],
  258. use_debugger=settings['server']['debug'],
  259. port=settings['server']['port']
  260. )
  261. if __name__ == "__main__":
  262. run()