webapp.py 16 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512
  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. if __name__ == '__main__':
  16. from sys import path
  17. from os.path import realpath, dirname
  18. path.append(realpath(dirname(realpath(__file__))+'/../'))
  19. import json
  20. import cStringIO
  21. import os
  22. from datetime import datetime, timedelta
  23. from itertools import chain
  24. from flask import (
  25. Flask, request, render_template, url_for, Response, make_response,
  26. redirect, send_from_directory
  27. )
  28. from flask.ext.babel import Babel, gettext, format_date
  29. from searx import settings, searx_dir
  30. from searx.engines import (
  31. categories, engines, get_engines_stats, engine_shortcuts
  32. )
  33. from searx.utils import (
  34. UnicodeWriter, highlight_content, html_to_text, get_themes
  35. )
  36. from searx.version import VERSION_STRING
  37. from searx.languages import language_codes
  38. from searx.https_rewrite import https_url_rewrite
  39. from searx.search import Search
  40. from searx.query import Query
  41. from searx.autocomplete import backends as autocomplete_backends
  42. static_path, templates_path, themes =\
  43. get_themes(settings['themes_path']
  44. if settings.get('themes_path')
  45. else searx_dir)
  46. default_theme = settings['server'].get('default_theme', 'default')
  47. app = Flask(
  48. __name__,
  49. static_folder=static_path,
  50. template_folder=templates_path
  51. )
  52. app.secret_key = settings['server']['secret_key']
  53. babel = Babel(app)
  54. #TODO configurable via settings.yml
  55. favicons = ['wikipedia', 'youtube', 'vimeo', 'dailymotion', 'soundcloud',
  56. 'twitter', 'stackoverflow', 'github', 'deviantart', 'kickass']
  57. cookie_max_age = 60 * 60 * 24 * 365 * 23 # 23 years
  58. @babel.localeselector
  59. def get_locale():
  60. locale = request.accept_languages.best_match(settings['locales'].keys())
  61. if request.cookies.get('locale', '') in settings['locales']:
  62. locale = request.cookies.get('locale', '')
  63. if 'locale' in request.args\
  64. and request.args['locale'] in settings['locales']:
  65. locale = request.args['locale']
  66. if 'locale' in request.form\
  67. and request.form['locale'] in settings['locales']:
  68. locale = request.form['locale']
  69. return locale
  70. def get_base_url():
  71. if settings['server']['base_url']:
  72. hostname = settings['server']['base_url']
  73. else:
  74. scheme = 'http'
  75. if request.is_secure:
  76. scheme = 'https'
  77. hostname = url_for('index', _external=True, _scheme=scheme)
  78. return hostname
  79. def get_current_theme_name(override=None):
  80. """Returns theme name.
  81. Checks in this order:
  82. 1. override
  83. 2. cookies
  84. 3. settings"""
  85. if override and override in themes:
  86. return override
  87. theme_name = request.args.get('theme',
  88. request.cookies.get('theme',
  89. default_theme))
  90. if theme_name not in themes:
  91. theme_name = default_theme
  92. return theme_name
  93. def url_for_theme(endpoint, override_theme=None, **values):
  94. if endpoint == 'static' and values.get('filename', None):
  95. theme_name = get_current_theme_name(override=override_theme)
  96. values['filename'] = "{}/{}".format(theme_name, values['filename'])
  97. return url_for(endpoint, **values)
  98. def render(template_name, override_theme=None, **kwargs):
  99. blocked_engines = request.cookies.get('blocked_engines', '').split(',')
  100. autocomplete = request.cookies.get('autocomplete')
  101. if autocomplete not in autocomplete_backends:
  102. autocomplete = None
  103. nonblocked_categories = (engines[e].categories
  104. for e in engines
  105. if e not in blocked_engines)
  106. nonblocked_categories = set(chain.from_iterable(nonblocked_categories))
  107. if 'categories' not in kwargs:
  108. kwargs['categories'] = ['general']
  109. kwargs['categories'].extend(x for x in
  110. sorted(categories.keys())
  111. if x != 'general'
  112. and x in nonblocked_categories)
  113. if 'selected_categories' not in kwargs:
  114. kwargs['selected_categories'] = []
  115. for arg in request.args:
  116. if arg.startswith('category_'):
  117. c = arg.split('_', 1)[1]
  118. if c in categories:
  119. kwargs['selected_categories'].append(c)
  120. if not kwargs['selected_categories']:
  121. cookie_categories = request.cookies.get('categories', '').split(',')
  122. for ccateg in cookie_categories:
  123. if ccateg in categories:
  124. kwargs['selected_categories'].append(ccateg)
  125. if not kwargs['selected_categories']:
  126. kwargs['selected_categories'] = ['general']
  127. if 'autocomplete' not in kwargs:
  128. kwargs['autocomplete'] = autocomplete
  129. kwargs['searx_version'] = VERSION_STRING
  130. kwargs['method'] = request.cookies.get('method', 'POST')
  131. # override url_for function in templates
  132. kwargs['url_for'] = url_for_theme
  133. kwargs['theme'] = get_current_theme_name(override=override_theme)
  134. kwargs['template_name'] = template_name
  135. return render_template(
  136. '{}/{}'.format(kwargs['theme'], template_name), **kwargs)
  137. @app.route('/search', methods=['GET', 'POST'])
  138. @app.route('/', methods=['GET', 'POST'])
  139. def index():
  140. """Render index page.
  141. Supported outputs: html, json, csv, rss.
  142. """
  143. if not request.args and not request.form:
  144. return render(
  145. 'index.html',
  146. )
  147. try:
  148. search = Search(request)
  149. except:
  150. return render(
  151. 'index.html',
  152. )
  153. search.results, search.suggestions,\
  154. search.answers, search.infoboxes = search.search(request)
  155. for result in search.results:
  156. if not search.paging and engines[result['engine']].paging:
  157. search.paging = True
  158. # check if HTTPS rewrite is required
  159. if settings['server']['https_rewrite']\
  160. and result['parsed_url'].scheme == 'http':
  161. result = https_url_rewrite(result)
  162. if search.request_data.get('format', 'html') == 'html':
  163. if 'content' in result:
  164. result['content'] = highlight_content(result['content'],
  165. search.query.encode('utf-8')) # noqa
  166. result['title'] = highlight_content(result['title'],
  167. search.query.encode('utf-8'))
  168. else:
  169. if 'content' in result:
  170. result['content'] = html_to_text(result['content']).strip()
  171. # removing html content and whitespace duplications
  172. result['title'] = ' '.join(html_to_text(result['title'])
  173. .strip().split())
  174. if len(result['url']) > 74:
  175. url_parts = result['url'][:35], result['url'][-35:]
  176. result['pretty_url'] = u'{0}[...]{1}'.format(*url_parts)
  177. else:
  178. result['pretty_url'] = result['url']
  179. for engine in result['engines']:
  180. if engine in favicons:
  181. result['favicon'] = engine
  182. # TODO, check if timezone is calculated right
  183. if 'publishedDate' in result:
  184. result['pubdate'] = result['publishedDate'].strftime('%Y-%m-%d %H:%M:%S%z')
  185. if result['publishedDate'].replace(tzinfo=None) >= datetime.now() - timedelta(days=1):
  186. timedifference = datetime.now() - result['publishedDate'].replace(tzinfo=None)
  187. minutes = int((timedifference.seconds / 60) % 60)
  188. hours = int(timedifference.seconds / 60 / 60)
  189. if hours == 0:
  190. result['publishedDate'] = gettext(u'{minutes} minute(s) ago').format(minutes=minutes) # noqa
  191. else:
  192. result['publishedDate'] = gettext(u'{hours} hour(s), {minutes} minute(s) ago').format(hours=hours, minutes=minutes) # noqa
  193. else:
  194. result['publishedDate'] = format_date(result['publishedDate'])
  195. if search.request_data.get('format') == 'json':
  196. return Response(json.dumps({'query': search.query,
  197. 'results': search.results}),
  198. mimetype='application/json')
  199. elif search.request_data.get('format') == 'csv':
  200. csv = UnicodeWriter(cStringIO.StringIO())
  201. keys = ('title', 'url', 'content', 'host', 'engine', 'score')
  202. if search.results:
  203. csv.writerow(keys)
  204. for row in search.results:
  205. row['host'] = row['parsed_url'].netloc
  206. csv.writerow([row.get(key, '') for key in keys])
  207. csv.stream.seek(0)
  208. response = Response(csv.stream.read(), mimetype='application/csv')
  209. cont_disp = 'attachment;Filename=searx_-_{0}.csv'.format(search.query)
  210. response.headers.add('Content-Disposition', cont_disp)
  211. return response
  212. elif search.request_data.get('format') == 'rss':
  213. response_rss = render(
  214. 'opensearch_response_rss.xml',
  215. results=search.results,
  216. q=search.request_data['q'],
  217. number_of_results=len(search.results),
  218. base_url=get_base_url()
  219. )
  220. return Response(response_rss, mimetype='text/xml')
  221. return render(
  222. 'results.html',
  223. results=search.results,
  224. q=search.request_data['q'],
  225. selected_categories=search.categories,
  226. paging=search.paging,
  227. pageno=search.pageno,
  228. base_url=get_base_url(),
  229. suggestions=search.suggestions,
  230. answers=search.answers,
  231. infoboxes=search.infoboxes,
  232. theme=get_current_theme_name()
  233. )
  234. @app.route('/about', methods=['GET'])
  235. def about():
  236. """Render about page"""
  237. return render(
  238. 'about.html',
  239. )
  240. @app.route('/autocompleter', methods=['GET', 'POST'])
  241. def autocompleter():
  242. """Return autocompleter results"""
  243. request_data = {}
  244. # select request method
  245. if request.method == 'POST':
  246. request_data = request.form
  247. else:
  248. request_data = request.args
  249. # set blocked engines
  250. if request.cookies.get('blocked_engines'):
  251. blocked_engines = request.cookies['blocked_engines'].split(',') # noqa
  252. else:
  253. blocked_engines = []
  254. # parse query
  255. query = Query(request_data.get('q', '').encode('utf-8'), blocked_engines)
  256. query.parse_query()
  257. # check if search query is set
  258. if not query.getSearchQuery():
  259. return
  260. # run autocompleter
  261. completer = autocomplete_backends.get(request.cookies.get('autocomplete'))
  262. # check if valid autocompleter is selected
  263. if not completer:
  264. return
  265. # run autocompletion
  266. raw_results = completer(query.getSearchQuery())
  267. # parse results (write :language and !engine back to result string)
  268. results = []
  269. for result in raw_results:
  270. result_query = query
  271. result_query.changeSearchQuery(result)
  272. # add parsed result
  273. results.append(result_query.getFullQuery())
  274. # return autocompleter results
  275. if request_data.get('format') == 'x-suggestions':
  276. return Response(json.dumps([query.query, results]),
  277. mimetype='application/json')
  278. else:
  279. return Response(json.dumps(results),
  280. mimetype='application/json')
  281. @app.route('/preferences', methods=['GET', 'POST'])
  282. def preferences():
  283. """Render preferences page.
  284. Settings that are going to be saved as cookies."""
  285. lang = None
  286. if request.cookies.get('language')\
  287. and request.cookies['language'] in (x[0] for x in language_codes):
  288. lang = request.cookies['language']
  289. blocked_engines = []
  290. if request.method == 'GET':
  291. blocked_engines = request.cookies.get('blocked_engines', '').split(',')
  292. else: # on save
  293. selected_categories = []
  294. locale = None
  295. autocomplete = ''
  296. method = 'POST'
  297. for pd_name, pd in request.form.items():
  298. if pd_name.startswith('category_'):
  299. category = pd_name[9:]
  300. if category not in categories:
  301. continue
  302. selected_categories.append(category)
  303. elif pd_name == 'locale' and pd in settings['locales']:
  304. locale = pd
  305. elif pd_name == 'autocomplete':
  306. autocomplete = pd
  307. elif pd_name == 'language' and (pd == 'all' or
  308. pd in (x[0] for
  309. x in language_codes)):
  310. lang = pd
  311. elif pd_name == 'method':
  312. method = pd
  313. elif pd_name.startswith('engine_'):
  314. engine_name = pd_name.replace('engine_', '', 1)
  315. if engine_name in engines:
  316. blocked_engines.append(engine_name)
  317. elif pd_name == 'theme':
  318. theme = pd if pd in themes else default_theme
  319. resp = make_response(redirect(url_for('index')))
  320. user_blocked_engines = request.cookies.get('blocked_engines', '').split(',') # noqa
  321. if sorted(blocked_engines) != sorted(user_blocked_engines):
  322. resp.set_cookie(
  323. 'blocked_engines', ','.join(blocked_engines),
  324. max_age=cookie_max_age
  325. )
  326. if locale:
  327. resp.set_cookie(
  328. 'locale', locale,
  329. max_age=cookie_max_age
  330. )
  331. if lang:
  332. resp.set_cookie(
  333. 'language', lang,
  334. max_age=cookie_max_age
  335. )
  336. if selected_categories:
  337. # cookie max age: 4 weeks
  338. resp.set_cookie(
  339. 'categories', ','.join(selected_categories),
  340. max_age=cookie_max_age
  341. )
  342. resp.set_cookie(
  343. 'autocomplete', autocomplete,
  344. max_age=cookie_max_age
  345. )
  346. resp.set_cookie('method', method, max_age=cookie_max_age)
  347. resp.set_cookie(
  348. 'theme', theme, max_age=cookie_max_age)
  349. return resp
  350. return render('preferences.html',
  351. locales=settings['locales'],
  352. current_locale=get_locale(),
  353. current_language=lang or 'all',
  354. language_codes=language_codes,
  355. categs=categories.items(),
  356. blocked_engines=blocked_engines,
  357. autocomplete_backends=autocomplete_backends,
  358. shortcuts={y: x for x, y in engine_shortcuts.items()},
  359. themes=themes,
  360. theme=get_current_theme_name())
  361. @app.route('/stats', methods=['GET'])
  362. def stats():
  363. """Render engine statistics page."""
  364. stats = get_engines_stats()
  365. return render(
  366. 'stats.html',
  367. stats=stats,
  368. )
  369. @app.route('/robots.txt', methods=['GET'])
  370. def robots():
  371. return Response("""User-agent: *
  372. Allow: /
  373. Allow: /about
  374. Disallow: /stats
  375. Disallow: /preferences
  376. """, mimetype='text/plain')
  377. @app.route('/opensearch.xml', methods=['GET'])
  378. def opensearch():
  379. method = 'post'
  380. # chrome/chromium only supports HTTP GET....
  381. if request.headers.get('User-Agent', '').lower().find('webkit') >= 0:
  382. method = 'get'
  383. ret = render('opensearch.xml',
  384. opensearch_method=method,
  385. host=get_base_url())
  386. resp = Response(response=ret,
  387. status=200,
  388. mimetype="application/xml")
  389. return resp
  390. @app.route('/favicon.ico')
  391. def favicon():
  392. return send_from_directory(os.path.join(app.root_path,
  393. 'static',
  394. get_current_theme_name(),
  395. 'img'),
  396. 'favicon.png',
  397. mimetype='image/vnd.microsoft.icon')
  398. def run():
  399. app.run(
  400. debug=settings['server']['debug'],
  401. use_debugger=settings['server']['debug'],
  402. port=settings['server']['port']
  403. )
  404. application = app
  405. if __name__ == "__main__":
  406. run()