webapp.py 17 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519
  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. from gevent import monkey
  16. monkey.patch_all()
  17. if __name__ == '__main__':
  18. from sys import path
  19. from os.path import realpath, dirname
  20. path.append(realpath(dirname(realpath(__file__))+'/../'))
  21. import json
  22. import cStringIO
  23. import os
  24. from datetime import datetime, timedelta
  25. from itertools import chain
  26. from flask import (
  27. Flask, request, render_template, url_for, Response, make_response,
  28. redirect, send_from_directory
  29. )
  30. from flask.ext.babel import Babel, gettext, format_date
  31. from searx import settings, searx_dir
  32. from searx.engines import (
  33. categories, engines, get_engines_stats, engine_shortcuts
  34. )
  35. from searx.utils import (
  36. UnicodeWriter, highlight_content, html_to_text, get_themes
  37. )
  38. from searx.https_rewrite import https_rules
  39. from searx.languages import language_codes
  40. from searx.search import Search
  41. from searx.query import Query
  42. from searx.autocomplete import backends as autocomplete_backends
  43. static_path, templates_path, themes =\
  44. get_themes(settings['themes_path']
  45. if settings.get('themes_path')
  46. else searx_dir)
  47. default_theme = settings['default_theme'] if \
  48. settings.get('default_theme', None) else 'default'
  49. app = Flask(
  50. __name__,
  51. static_folder=static_path,
  52. template_folder=templates_path
  53. )
  54. app.secret_key = settings['server']['secret_key']
  55. babel = Babel(app)
  56. #TODO configurable via settings.yml
  57. favicons = ['wikipedia', 'youtube', 'vimeo', 'soundcloud',
  58. 'twitter', 'stackoverflow', 'github']
  59. cookie_max_age = 60 * 60 * 24 * 365 * 23 # 23 years
  60. @babel.localeselector
  61. def get_locale():
  62. locale = request.accept_languages.best_match(settings['locales'].keys())
  63. if request.cookies.get('locale', '') in settings['locales']:
  64. locale = request.cookies.get('locale', '')
  65. if 'locale' in request.args\
  66. and request.args['locale'] in settings['locales']:
  67. locale = request.args['locale']
  68. if 'locale' in request.form\
  69. and request.form['locale'] in settings['locales']:
  70. locale = request.form['locale']
  71. return locale
  72. def get_base_url():
  73. if settings['server']['base_url']:
  74. hostname = settings['server']['base_url']
  75. else:
  76. scheme = 'http'
  77. if request.is_secure:
  78. scheme = 'https'
  79. hostname = url_for('index', _external=True, _scheme=scheme)
  80. return hostname
  81. def get_current_theme_name(override=None):
  82. """Returns theme name.
  83. Checks in this order:
  84. 1. override
  85. 2. cookies
  86. 3. settings"""
  87. if override and override in themes:
  88. return override
  89. theme_name = request.args.get('theme',
  90. request.cookies.get('theme',
  91. default_theme))
  92. if theme_name not in themes:
  93. theme_name = default_theme
  94. return theme_name
  95. def url_for_theme(endpoint, override_theme=None, **values):
  96. if endpoint == 'static' and values.get('filename', None):
  97. theme_name = get_current_theme_name(override=override_theme)
  98. values['filename'] = "{}/{}".format(theme_name, values['filename'])
  99. return url_for(endpoint, **values)
  100. def render(template_name, override_theme=None, **kwargs):
  101. blocked_engines = request.cookies.get('blocked_engines', '').split(',')
  102. autocomplete = request.cookies.get('autocomplete')
  103. if autocomplete not in autocomplete_backends:
  104. autocomplete = None
  105. nonblocked_categories = (engines[e].categories
  106. for e in engines
  107. if e not in blocked_engines)
  108. nonblocked_categories = set(chain.from_iterable(nonblocked_categories))
  109. if not 'categories' in kwargs:
  110. kwargs['categories'] = ['general']
  111. kwargs['categories'].extend(x for x in
  112. sorted(categories.keys())
  113. if x != 'general'
  114. and x in nonblocked_categories)
  115. if not 'selected_categories' in kwargs:
  116. kwargs['selected_categories'] = []
  117. for arg in request.args:
  118. if arg.startswith('category_'):
  119. c = arg.split('_', 1)[1]
  120. if c in categories:
  121. kwargs['selected_categories'].append(c)
  122. if not kwargs['selected_categories']:
  123. cookie_categories = request.cookies.get('categories', '').split(',')
  124. for ccateg in cookie_categories:
  125. if ccateg in categories:
  126. kwargs['selected_categories'].append(ccateg)
  127. if not kwargs['selected_categories']:
  128. kwargs['selected_categories'] = ['general']
  129. if not 'autocomplete' in kwargs:
  130. kwargs['autocomplete'] = autocomplete
  131. kwargs['method'] = request.cookies.get('method', 'POST')
  132. # override url_for function in templates
  133. kwargs['url_for'] = url_for_theme
  134. kwargs['theme'] = get_current_theme_name(override=override_theme)
  135. kwargs['template_name'] = template_name
  136. return render_template(
  137. '{}/{}'.format(kwargs['theme'], template_name), **kwargs)
  138. @app.route('/search', methods=['GET', 'POST'])
  139. @app.route('/', methods=['GET', 'POST'])
  140. def index():
  141. """Render index page.
  142. Supported outputs: html, json, csv, rss.
  143. """
  144. if not request.args and not request.form:
  145. return render(
  146. 'index.html',
  147. )
  148. try:
  149. search = Search(request)
  150. except:
  151. return render(
  152. 'index.html',
  153. )
  154. search.results, search.suggestions, 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. if settings['server']['https_rewrite']\
  159. and result['parsed_url'].scheme == 'http':
  160. for http_regex, https_url in https_rules:
  161. if http_regex.match(result['url']):
  162. result['url'] = http_regex.sub(https_url, result['url'])
  163. # TODO result['parsed_url'].scheme
  164. break
  165. # HTTPS rewrite
  166. if search.request_data.get('format', 'html') == 'html':
  167. if 'content' in result:
  168. result['content'] = highlight_content(result['content'],
  169. search.query.encode('utf-8')) # noqa
  170. result['title'] = highlight_content(result['title'],
  171. search.query.encode('utf-8'))
  172. else:
  173. if 'content' in result:
  174. result['content'] = html_to_text(result['content']).strip()
  175. # removing html content and whitespace duplications
  176. result['title'] = ' '.join(html_to_text(result['title'])
  177. .strip().split())
  178. if len(result['url']) > 74:
  179. url_parts = result['url'][:35], result['url'][-35:]
  180. result['pretty_url'] = u'{0}[...]{1}'.format(*url_parts)
  181. else:
  182. result['pretty_url'] = result['url']
  183. for engine in result['engines']:
  184. if engine in favicons:
  185. result['favicon'] = engine
  186. # TODO, check if timezone is calculated right
  187. if 'publishedDate' in result:
  188. if result['publishedDate'].replace(tzinfo=None)\
  189. >= datetime.now() - timedelta(days=1):
  190. timedifference = datetime.now() - result['publishedDate']\
  191. .replace(tzinfo=None)
  192. minutes = int((timedifference.seconds / 60) % 60)
  193. hours = int(timedifference.seconds / 60 / 60)
  194. if hours == 0:
  195. result['publishedDate'] = gettext(u'{minutes} minute(s) ago').format(minutes=minutes) # noqa
  196. else:
  197. result['publishedDate'] = gettext(u'{hours} hour(s), {minutes} minute(s) ago').format(hours=hours, minutes=minutes) # noqa
  198. else:
  199. result['pubdate'] = result['publishedDate']\
  200. .strftime('%a, %d %b %Y %H:%M:%S %z')
  201. result['publishedDate'] = format_date(result['publishedDate'])
  202. if search.request_data.get('format') == 'json':
  203. return Response(json.dumps({'query': search.query,
  204. 'results': search.results}),
  205. mimetype='application/json')
  206. elif search.request_data.get('format') == 'csv':
  207. csv = UnicodeWriter(cStringIO.StringIO())
  208. keys = ('title', 'url', 'content', 'host', 'engine', 'score')
  209. if search.results:
  210. csv.writerow(keys)
  211. for row in search.results:
  212. row['host'] = row['parsed_url'].netloc
  213. csv.writerow([row.get(key, '') for key in keys])
  214. csv.stream.seek(0)
  215. response = Response(csv.stream.read(), mimetype='application/csv')
  216. cont_disp = 'attachment;Filename=searx_-_{0}.csv'.format(search.query)
  217. response.headers.add('Content-Disposition', cont_disp)
  218. return response
  219. elif search.request_data.get('format') == 'rss':
  220. response_rss = render(
  221. 'opensearch_response_rss.xml',
  222. results=search.results,
  223. q=search.request_data['q'],
  224. number_of_results=len(search.results),
  225. base_url=get_base_url()
  226. )
  227. return Response(response_rss, mimetype='text/xml')
  228. return render(
  229. 'results.html',
  230. results=search.results,
  231. q=search.request_data['q'],
  232. selected_categories=search.categories,
  233. paging=search.paging,
  234. pageno=search.pageno,
  235. base_url=get_base_url(),
  236. suggestions=search.suggestions,
  237. answers=search.answers,
  238. infoboxes=search.infoboxes,
  239. theme=get_current_theme_name()
  240. )
  241. @app.route('/about', methods=['GET'])
  242. def about():
  243. """Render about page"""
  244. return render(
  245. 'about.html',
  246. )
  247. @app.route('/autocompleter', methods=['GET', 'POST'])
  248. def autocompleter():
  249. """Return autocompleter results"""
  250. request_data = {}
  251. # select request method
  252. if request.method == 'POST':
  253. request_data = request.form
  254. else:
  255. request_data = request.args
  256. # set blocked engines
  257. if request.cookies.get('blocked_engines'):
  258. blocked_engines = request.cookies['blocked_engines'].split(',') # noqa
  259. else:
  260. blocked_engines = []
  261. # parse query
  262. query = Query(request_data.get('q', '').encode('utf-8'), blocked_engines)
  263. query.parse_query()
  264. # check if search query is set
  265. if not query.getSearchQuery():
  266. return
  267. # run autocompleter
  268. completer = autocomplete_backends.get(request.cookies.get('autocomplete'))
  269. # check if valid autocompleter is selected
  270. if not completer:
  271. return
  272. # run autocompletion
  273. raw_results = completer(query.getSearchQuery())
  274. # parse results (write :language and !engine back to result string)
  275. results = []
  276. for result in raw_results:
  277. result_query = query
  278. result_query.changeSearchQuery(result)
  279. # add parsed result
  280. results.append(result_query.getFullQuery())
  281. # return autocompleter results
  282. if request_data.get('format') == 'x-suggestions':
  283. return Response(json.dumps([query, results]),
  284. mimetype='application/json')
  285. else:
  286. return Response(json.dumps(results),
  287. mimetype='application/json')
  288. @app.route('/preferences', methods=['GET', 'POST'])
  289. def preferences():
  290. """Render preferences page.
  291. Settings that are going to be saved as cookies."""
  292. lang = None
  293. if request.cookies.get('language')\
  294. and request.cookies['language'] in (x[0] for x in language_codes):
  295. lang = request.cookies['language']
  296. blocked_engines = []
  297. if request.method == 'GET':
  298. blocked_engines = request.cookies.get('blocked_engines', '').split(',')
  299. else: # on save
  300. selected_categories = []
  301. locale = None
  302. autocomplete = ''
  303. method = 'POST'
  304. for pd_name, pd in request.form.items():
  305. if pd_name.startswith('category_'):
  306. category = pd_name[9:]
  307. if not category in categories:
  308. continue
  309. selected_categories.append(category)
  310. elif pd_name == 'locale' and pd in settings['locales']:
  311. locale = pd
  312. elif pd_name == 'autocomplete':
  313. autocomplete = pd
  314. elif pd_name == 'language' and (pd == 'all' or
  315. pd in (x[0] for
  316. x in language_codes)):
  317. lang = pd
  318. elif pd_name == 'method':
  319. method = pd
  320. elif pd_name.startswith('engine_'):
  321. engine_name = pd_name.replace('engine_', '', 1)
  322. if engine_name in engines:
  323. blocked_engines.append(engine_name)
  324. elif pd_name == 'theme':
  325. theme = pd if pd in themes else default_theme
  326. resp = make_response(redirect(url_for('index')))
  327. user_blocked_engines = request.cookies.get('blocked_engines', '').split(',') # noqa
  328. if sorted(blocked_engines) != sorted(user_blocked_engines):
  329. resp.set_cookie(
  330. 'blocked_engines', ','.join(blocked_engines),
  331. max_age=cookie_max_age
  332. )
  333. if locale:
  334. resp.set_cookie(
  335. 'locale', locale,
  336. max_age=cookie_max_age
  337. )
  338. if lang:
  339. resp.set_cookie(
  340. 'language', lang,
  341. max_age=cookie_max_age
  342. )
  343. if selected_categories:
  344. # cookie max age: 4 weeks
  345. resp.set_cookie(
  346. 'categories', ','.join(selected_categories),
  347. max_age=cookie_max_age
  348. )
  349. resp.set_cookie(
  350. 'autocomplete', autocomplete,
  351. max_age=cookie_max_age
  352. )
  353. resp.set_cookie('method', method, max_age=cookie_max_age)
  354. resp.set_cookie(
  355. 'theme', theme, max_age=cookie_max_age)
  356. return resp
  357. return render('preferences.html',
  358. locales=settings['locales'],
  359. current_locale=get_locale(),
  360. current_language=lang or 'all',
  361. language_codes=language_codes,
  362. categs=categories.items(),
  363. blocked_engines=blocked_engines,
  364. autocomplete_backends=autocomplete_backends,
  365. shortcuts={y: x for x, y in engine_shortcuts.items()},
  366. themes=themes,
  367. theme=get_current_theme_name())
  368. @app.route('/stats', methods=['GET'])
  369. def stats():
  370. """Render engine statistics page."""
  371. stats = get_engines_stats()
  372. return render(
  373. 'stats.html',
  374. stats=stats,
  375. )
  376. @app.route('/robots.txt', methods=['GET'])
  377. def robots():
  378. return Response("""User-agent: *
  379. Allow: /
  380. Allow: /about
  381. Disallow: /stats
  382. Disallow: /preferences
  383. """, mimetype='text/plain')
  384. @app.route('/opensearch.xml', methods=['GET'])
  385. def opensearch():
  386. method = 'post'
  387. # chrome/chromium only supports HTTP GET....
  388. if request.headers.get('User-Agent', '').lower().find('webkit') >= 0:
  389. method = 'get'
  390. ret = render('opensearch.xml',
  391. opensearch_method=method,
  392. host=get_base_url())
  393. resp = Response(response=ret,
  394. status=200,
  395. mimetype="application/xml")
  396. return resp
  397. @app.route('/favicon.ico')
  398. def favicon():
  399. return send_from_directory(os.path.join(app.root_path,
  400. 'static',
  401. get_current_theme_name(),
  402. 'img'),
  403. 'favicon.png',
  404. mimetype='image/vnd.microsoft.icon')
  405. def run():
  406. app.run(
  407. debug=settings['server']['debug'],
  408. use_debugger=settings['server']['debug'],
  409. port=settings['server']['port']
  410. )
  411. application = app
  412. if __name__ == "__main__":
  413. run()