webapp.py 10 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318
  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. @babel.localeselector
  37. def get_locale():
  38. locale = request.accept_languages.best_match(settings['locales'].keys())
  39. if request.cookies.get('locale', '') in settings['locales']:
  40. locale = request.cookies.get('locale', '')
  41. if 'locale' in request.args\
  42. and request.args['locale'] in settings['locales']:
  43. locale = request.args['locale']
  44. if 'locale' in request.form\
  45. and request.form['locale'] in settings['locales']:
  46. locale = request.form['locale']
  47. return locale
  48. def get_base_url():
  49. if settings['server']['base_url']:
  50. hostname = settings['server']['base_url']
  51. else:
  52. scheme = 'http'
  53. if request.is_secure:
  54. scheme = 'https'
  55. hostname = url_for('index', _external=True, _scheme=scheme)
  56. return hostname
  57. def render(template_name, **kwargs):
  58. global categories
  59. kwargs['categories'] = ['general']
  60. kwargs['categories'].extend(x for x in
  61. sorted(categories.keys()) if x != 'general')
  62. if not 'selected_categories' in kwargs:
  63. kwargs['selected_categories'] = []
  64. cookie_categories = request.cookies.get('categories', '').split(',')
  65. for ccateg in cookie_categories:
  66. if ccateg in categories:
  67. kwargs['selected_categories'].append(ccateg)
  68. if not len(kwargs['selected_categories']):
  69. kwargs['selected_categories'] = ['general']
  70. return render_template(template_name, **kwargs)
  71. def parse_query(query):
  72. query_engines = []
  73. query_parts = query.split()
  74. if query_parts[0].startswith('-'):
  75. engine_name = query_parts[0][1:].replace('_', ' ')
  76. if engine_name in engines:
  77. query_engines.append({'category': 'none',
  78. 'name': query_parts[0][1:]})
  79. elif query_parts[0].startswith('!'):
  80. category_name = query_parts[0][1:].replace('_', ' ')
  81. if category_name in categories:
  82. query_engines.extend({'category': category_name,
  83. 'name': engine.name}
  84. for engine in categories[category_name])
  85. if len(query_engines):
  86. query = query.replace(query_parts[0], '', 1).strip()
  87. return query, query_engines
  88. @app.route('/', methods=['GET', 'POST'])
  89. def index():
  90. paging = False
  91. if request.method == 'POST':
  92. request_data = request.form
  93. else:
  94. request_data = request.args
  95. if not request_data.get('q'):
  96. return render('index.html')
  97. pageno_param = request_data.get('pageno', '1')
  98. if not pageno_param.isdigit() or int(pageno_param) < 1:
  99. return render('index.html')
  100. pageno = int(pageno_param)
  101. selected_categories = []
  102. query, selected_engines = parse_query(request_data['q'].encode('utf-8'))
  103. if len(selected_engines):
  104. selected_categories = list(set(engine['category']
  105. for engine in selected_engines))
  106. else:
  107. for pd_name, pd in request_data.items():
  108. if pd_name.startswith('category_'):
  109. category = pd_name[9:]
  110. if not category in categories:
  111. continue
  112. selected_categories.append(category)
  113. if not len(selected_categories):
  114. cookie_categories = request.cookies.get('categories', '')
  115. cookie_categories = cookie_categories.split(',')
  116. for ccateg in cookie_categories:
  117. if ccateg in categories:
  118. selected_categories.append(ccateg)
  119. if not len(selected_categories):
  120. selected_categories = ['general']
  121. for categ in selected_categories:
  122. selected_engines.extend({'category': categ,
  123. 'name': x.name}
  124. for x in categories[categ])
  125. results, suggestions = search(query, request, selected_engines, pageno)
  126. for result in results:
  127. if not paging and engines[result['engine']].paging:
  128. paging = True
  129. if request_data.get('format', 'html') == 'html':
  130. if 'content' in result:
  131. result['content'] = highlight_content(result['content'], query)
  132. result['title'] = highlight_content(result['title'], query)
  133. else:
  134. if 'content' in result:
  135. result['content'] = html_to_text(result['content']).strip()
  136. result['title'] = html_to_text(result['title']).strip()
  137. if len(result['url']) > 74:
  138. url_parts = result['url'][:35], result['url'][-35:]
  139. result['pretty_url'] = '{0}[...]{1}'.format(*url_parts)
  140. else:
  141. result['pretty_url'] = result['url']
  142. for engine in result['engines']:
  143. if engine in favicons:
  144. result['favicon'] = engine
  145. if request_data.get('format') == 'json':
  146. return Response(json.dumps({'query': query, 'results': results}),
  147. mimetype='application/json')
  148. elif request_data.get('format') == 'csv':
  149. csv = UnicodeWriter(cStringIO.StringIO())
  150. keys = ('title', 'url', 'content', 'host', 'engine', 'score')
  151. if len(results):
  152. csv.writerow(keys)
  153. for row in results:
  154. row['host'] = row['parsed_url'].netloc
  155. csv.writerow([row.get(key, '') for key in keys])
  156. csv.stream.seek(0)
  157. response = Response(csv.stream.read(), mimetype='application/csv')
  158. content_disp = 'attachment;Filename=searx_-_{0}.csv'.format(query)
  159. response.headers.add('Content-Disposition', content_disp)
  160. return response
  161. elif request_data.get('format') == 'rss':
  162. response_rss = render(
  163. 'opensearch_response_rss.xml',
  164. results=results,
  165. q=request_data['q'],
  166. number_of_results=len(results),
  167. base_url=get_base_url()
  168. )
  169. return Response(response_rss, mimetype='text/xml')
  170. return render(
  171. 'results.html',
  172. results=results,
  173. q=request_data['q'],
  174. selected_categories=selected_categories,
  175. paging=paging,
  176. pageno=pageno,
  177. suggestions=suggestions
  178. )
  179. @app.route('/about', methods=['GET'])
  180. def about():
  181. return render('about.html')
  182. @app.route('/engines', methods=['GET'])
  183. def list_engines():
  184. global categories
  185. return render('engines.html', categs=categories.items())
  186. @app.route('/preferences', methods=['GET', 'POST'])
  187. def preferences():
  188. if request.method == 'POST':
  189. selected_categories = []
  190. locale = None
  191. for pd_name, pd in request.form.items():
  192. if pd_name.startswith('category_'):
  193. category = pd_name[9:]
  194. if not category in categories:
  195. continue
  196. selected_categories.append(category)
  197. elif pd_name == 'locale' and pd in settings['locales']:
  198. locale = pd
  199. resp = make_response(redirect('/'))
  200. if locale:
  201. # cookie max age: 4 weeks
  202. resp.set_cookie(
  203. 'locale', locale,
  204. max_age=60 * 60 * 24 * 7 * 4
  205. )
  206. if selected_categories:
  207. # cookie max age: 4 weeks
  208. resp.set_cookie(
  209. 'categories', ','.join(selected_categories),
  210. max_age=60 * 60 * 24 * 7 * 4
  211. )
  212. return resp
  213. return render('preferences.html',
  214. locales=settings['locales'],
  215. current_locale=get_locale())
  216. @app.route('/stats', methods=['GET'])
  217. def stats():
  218. global categories
  219. stats = get_engines_stats()
  220. return render('stats.html', stats=stats)
  221. @app.route('/robots.txt', methods=['GET'])
  222. def robots():
  223. return Response("""User-agent: *
  224. Allow: /
  225. Allow: /about
  226. Disallow: /stats
  227. Disallow: /engines
  228. """, mimetype='text/plain')
  229. @app.route('/opensearch.xml', methods=['GET'])
  230. def opensearch():
  231. method = 'post'
  232. # chrome/chromium only supports HTTP GET....
  233. if request.headers.get('User-Agent', '').lower().find('webkit') >= 0:
  234. method = 'get'
  235. ret = render('opensearch.xml', method=method, host=get_base_url())
  236. resp = Response(response=ret,
  237. status=200,
  238. mimetype="application/xml")
  239. return resp
  240. @app.route('/favicon.ico')
  241. def favicon():
  242. return send_from_directory(os.path.join(app.root_path, 'static/img'),
  243. 'favicon.png',
  244. mimetype='image/vnd.microsoft.icon')
  245. def run():
  246. from gevent import monkey
  247. monkey.patch_all()
  248. app.run(
  249. debug=settings['server']['debug'],
  250. use_debugger=settings['server']['debug'],
  251. port=settings['server']['port']
  252. )
  253. if __name__ == "__main__":
  254. run()