webapp.py 17 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517
  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. get_static_files
  36. )
  37. from searx.version import VERSION_STRING
  38. from searx.languages import language_codes
  39. from searx.https_rewrite import https_url_rewrite
  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['server'].get('default_theme', 'default')
  48. static_files = get_static_files(searx_dir)
  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. global_favicons = []
  57. for indice, theme in enumerate(themes):
  58. global_favicons.append([])
  59. theme_img_path = searx_dir+"/static/"+theme+"/img/icons/"
  60. for (dirpath, dirnames, filenames) in os.walk(theme_img_path):
  61. global_favicons[indice].extend(filenames)
  62. cookie_max_age = 60 * 60 * 24 * 365 * 23 # 23 years
  63. @babel.localeselector
  64. def get_locale():
  65. locale = request.accept_languages.best_match(settings['locales'].keys())
  66. if request.cookies.get('locale', '') in settings['locales']:
  67. locale = request.cookies.get('locale', '')
  68. if 'locale' in request.args\
  69. and request.args['locale'] in settings['locales']:
  70. locale = request.args['locale']
  71. if 'locale' in request.form\
  72. and request.form['locale'] in settings['locales']:
  73. locale = request.form['locale']
  74. return locale
  75. def get_base_url():
  76. if settings['server']['base_url']:
  77. hostname = settings['server']['base_url']
  78. else:
  79. scheme = 'http'
  80. if request.is_secure:
  81. scheme = 'https'
  82. hostname = url_for('index', _external=True, _scheme=scheme)
  83. return hostname
  84. def get_current_theme_name(override=None):
  85. """Returns theme name.
  86. Checks in this order:
  87. 1. override
  88. 2. cookies
  89. 3. settings"""
  90. if override and override in themes:
  91. return override
  92. theme_name = request.args.get('theme',
  93. request.cookies.get('theme',
  94. default_theme))
  95. if theme_name not in themes:
  96. theme_name = default_theme
  97. return theme_name
  98. def url_for_theme(endpoint, override_theme=None, **values):
  99. if endpoint == 'static' and values.get('filename'):
  100. theme_name = get_current_theme_name(override=override_theme)
  101. filename_with_theme = "themes/{}/{}".format(theme_name, values['filename'])
  102. if filename_with_theme in static_files:
  103. values['filename'] = filename_with_theme
  104. return url_for(endpoint, **values)
  105. def render(template_name, override_theme=None, **kwargs):
  106. blocked_engines = request.cookies.get('blocked_engines', '').split(',')
  107. autocomplete = request.cookies.get('autocomplete')
  108. if autocomplete not in autocomplete_backends:
  109. autocomplete = None
  110. nonblocked_categories = (engines[e].categories
  111. for e in engines
  112. if e not in blocked_engines)
  113. nonblocked_categories = set(chain.from_iterable(nonblocked_categories))
  114. if 'categories' not in kwargs:
  115. kwargs['categories'] = ['general']
  116. kwargs['categories'].extend(x for x in
  117. sorted(categories.keys())
  118. if x != 'general'
  119. and x in nonblocked_categories)
  120. if 'selected_categories' not in kwargs:
  121. kwargs['selected_categories'] = []
  122. for arg in request.args:
  123. if arg.startswith('category_'):
  124. c = arg.split('_', 1)[1]
  125. if c in categories:
  126. kwargs['selected_categories'].append(c)
  127. if not kwargs['selected_categories']:
  128. cookie_categories = request.cookies.get('categories', '').split(',')
  129. for ccateg in cookie_categories:
  130. if ccateg in categories:
  131. kwargs['selected_categories'].append(ccateg)
  132. if not kwargs['selected_categories']:
  133. kwargs['selected_categories'] = ['general']
  134. if 'autocomplete' not in kwargs:
  135. kwargs['autocomplete'] = autocomplete
  136. kwargs['searx_version'] = VERSION_STRING
  137. kwargs['method'] = request.cookies.get('method', 'POST')
  138. # override url_for function in templates
  139. kwargs['url_for'] = url_for_theme
  140. kwargs['theme'] = get_current_theme_name(override=override_theme)
  141. kwargs['template_name'] = template_name
  142. return render_template(
  143. '{}/{}'.format(kwargs['theme'], template_name), **kwargs)
  144. @app.route('/search', methods=['GET', 'POST'])
  145. @app.route('/', methods=['GET', 'POST'])
  146. def index():
  147. """Render index page.
  148. Supported outputs: html, json, csv, rss.
  149. """
  150. if not request.args and not request.form:
  151. return render(
  152. 'index.html',
  153. )
  154. try:
  155. search = Search(request)
  156. except:
  157. return render(
  158. 'index.html',
  159. )
  160. search.results, search.suggestions,\
  161. search.answers, search.infoboxes = search.search(request)
  162. for result in search.results:
  163. if not search.paging and engines[result['engine']].paging:
  164. search.paging = True
  165. # check if HTTPS rewrite is required
  166. if settings['server']['https_rewrite']\
  167. and result['parsed_url'].scheme == 'http':
  168. result = https_url_rewrite(result)
  169. if search.request_data.get('format', 'html') == 'html':
  170. if 'content' in result:
  171. result['content'] = highlight_content(result['content'],
  172. search.query.encode('utf-8')) # noqa
  173. result['title'] = highlight_content(result['title'],
  174. search.query.encode('utf-8'))
  175. else:
  176. if 'content' in result:
  177. result['content'] = html_to_text(result['content']).strip()
  178. # removing html content and whitespace duplications
  179. result['title'] = ' '.join(html_to_text(result['title'])
  180. .strip().split())
  181. if len(result['url']) > 74:
  182. url_parts = result['url'][:35], result['url'][-35:]
  183. result['pretty_url'] = u'{0}[...]{1}'.format(*url_parts)
  184. else:
  185. result['pretty_url'] = result['url']
  186. # TODO, check if timezone is calculated right
  187. if 'publishedDate' in result:
  188. result['pubdate'] = result['publishedDate'].strftime('%Y-%m-%d %H:%M:%S%z')
  189. if result['publishedDate'].replace(tzinfo=None) >= datetime.now() - timedelta(days=1):
  190. timedifference = datetime.now() - result['publishedDate'].replace(tzinfo=None)
  191. minutes = int((timedifference.seconds / 60) % 60)
  192. hours = int(timedifference.seconds / 60 / 60)
  193. if hours == 0:
  194. result['publishedDate'] = gettext(u'{minutes} minute(s) ago').format(minutes=minutes) # noqa
  195. else:
  196. result['publishedDate'] = gettext(u'{hours} hour(s), {minutes} minute(s) ago').format(hours=hours, minutes=minutes) # noqa
  197. else:
  198. result['publishedDate'] = format_date(result['publishedDate'])
  199. if search.request_data.get('format') == 'json':
  200. return Response(json.dumps({'query': search.query,
  201. 'results': search.results}),
  202. mimetype='application/json')
  203. elif search.request_data.get('format') == 'csv':
  204. csv = UnicodeWriter(cStringIO.StringIO())
  205. keys = ('title', 'url', 'content', 'host', 'engine', 'score')
  206. if search.results:
  207. csv.writerow(keys)
  208. for row in search.results:
  209. row['host'] = row['parsed_url'].netloc
  210. csv.writerow([row.get(key, '') for key in keys])
  211. csv.stream.seek(0)
  212. response = Response(csv.stream.read(), mimetype='application/csv')
  213. cont_disp = 'attachment;Filename=searx_-_{0}.csv'.format(search.query)
  214. response.headers.add('Content-Disposition', cont_disp)
  215. return response
  216. elif search.request_data.get('format') == 'rss':
  217. response_rss = render(
  218. 'opensearch_response_rss.xml',
  219. results=search.results,
  220. q=search.request_data['q'],
  221. number_of_results=len(search.results),
  222. base_url=get_base_url()
  223. )
  224. return Response(response_rss, mimetype='text/xml')
  225. return render(
  226. 'results.html',
  227. results=search.results,
  228. q=search.request_data['q'],
  229. selected_categories=search.categories,
  230. paging=search.paging,
  231. pageno=search.pageno,
  232. base_url=get_base_url(),
  233. suggestions=search.suggestions,
  234. answers=search.answers,
  235. infoboxes=search.infoboxes,
  236. theme=get_current_theme_name(),
  237. favicons=global_favicons[themes.index(get_current_theme_name())]
  238. )
  239. @app.route('/about', methods=['GET'])
  240. def about():
  241. """Render about page"""
  242. return render(
  243. 'about.html',
  244. )
  245. @app.route('/autocompleter', methods=['GET', 'POST'])
  246. def autocompleter():
  247. """Return autocompleter results"""
  248. request_data = {}
  249. # select request method
  250. if request.method == 'POST':
  251. request_data = request.form
  252. else:
  253. request_data = request.args
  254. # set blocked engines
  255. if request.cookies.get('blocked_engines'):
  256. blocked_engines = request.cookies['blocked_engines'].split(',') # noqa
  257. else:
  258. blocked_engines = []
  259. # parse query
  260. query = Query(request_data.get('q', '').encode('utf-8'), blocked_engines)
  261. query.parse_query()
  262. # check if search query is set
  263. if not query.getSearchQuery():
  264. return
  265. # run autocompleter
  266. completer = autocomplete_backends.get(request.cookies.get('autocomplete'))
  267. # check if valid autocompleter is selected
  268. if not completer:
  269. return
  270. # run autocompletion
  271. raw_results = completer(query.getSearchQuery())
  272. # parse results (write :language and !engine back to result string)
  273. results = []
  274. for result in raw_results:
  275. result_query = query
  276. result_query.changeSearchQuery(result)
  277. # add parsed result
  278. results.append(result_query.getFullQuery())
  279. # return autocompleter results
  280. if request_data.get('format') == 'x-suggestions':
  281. return Response(json.dumps([query.query, results]),
  282. mimetype='application/json')
  283. else:
  284. return Response(json.dumps(results),
  285. mimetype='application/json')
  286. @app.route('/preferences', methods=['GET', 'POST'])
  287. def preferences():
  288. """Render preferences page.
  289. Settings that are going to be saved as cookies."""
  290. lang = None
  291. if request.cookies.get('language')\
  292. and request.cookies['language'] in (x[0] for x in language_codes):
  293. lang = request.cookies['language']
  294. blocked_engines = []
  295. if request.method == 'GET':
  296. blocked_engines = request.cookies.get('blocked_engines', '').split(',')
  297. else: # on save
  298. selected_categories = []
  299. locale = None
  300. autocomplete = ''
  301. method = 'POST'
  302. for pd_name, pd in request.form.items():
  303. if pd_name.startswith('category_'):
  304. category = pd_name[9:]
  305. if category not in categories:
  306. continue
  307. selected_categories.append(category)
  308. elif pd_name == 'locale' and pd in settings['locales']:
  309. locale = pd
  310. elif pd_name == 'autocomplete':
  311. autocomplete = pd
  312. elif pd_name == 'language' and (pd == 'all' or
  313. pd in (x[0] for
  314. x in language_codes)):
  315. lang = pd
  316. elif pd_name == 'method':
  317. method = pd
  318. elif pd_name.startswith('engine_'):
  319. engine_name = pd_name.replace('engine_', '', 1)
  320. if engine_name in engines:
  321. blocked_engines.append(engine_name)
  322. elif pd_name == 'theme':
  323. theme = pd if pd in themes else default_theme
  324. resp = make_response(redirect(url_for('index')))
  325. user_blocked_engines = request.cookies.get('blocked_engines', '').split(',') # noqa
  326. if sorted(blocked_engines) != sorted(user_blocked_engines):
  327. resp.set_cookie(
  328. 'blocked_engines', ','.join(blocked_engines),
  329. max_age=cookie_max_age
  330. )
  331. if locale:
  332. resp.set_cookie(
  333. 'locale', locale,
  334. max_age=cookie_max_age
  335. )
  336. if lang:
  337. resp.set_cookie(
  338. 'language', lang,
  339. max_age=cookie_max_age
  340. )
  341. if selected_categories:
  342. # cookie max age: 4 weeks
  343. resp.set_cookie(
  344. 'categories', ','.join(selected_categories),
  345. max_age=cookie_max_age
  346. )
  347. resp.set_cookie(
  348. 'autocomplete', autocomplete,
  349. max_age=cookie_max_age
  350. )
  351. resp.set_cookie('method', method, max_age=cookie_max_age)
  352. resp.set_cookie(
  353. 'theme', theme, max_age=cookie_max_age)
  354. return resp
  355. return render('preferences.html',
  356. locales=settings['locales'],
  357. current_locale=get_locale(),
  358. current_language=lang or 'all',
  359. language_codes=language_codes,
  360. categs=categories.items(),
  361. blocked_engines=blocked_engines,
  362. autocomplete_backends=autocomplete_backends,
  363. shortcuts={y: x for x, y in engine_shortcuts.items()},
  364. themes=themes,
  365. theme=get_current_theme_name())
  366. @app.route('/stats', methods=['GET'])
  367. def stats():
  368. """Render engine statistics page."""
  369. stats = get_engines_stats()
  370. return render(
  371. 'stats.html',
  372. stats=stats,
  373. )
  374. @app.route('/robots.txt', methods=['GET'])
  375. def robots():
  376. return Response("""User-agent: *
  377. Allow: /
  378. Allow: /about
  379. Disallow: /stats
  380. Disallow: /preferences
  381. """, mimetype='text/plain')
  382. @app.route('/opensearch.xml', methods=['GET'])
  383. def opensearch():
  384. method = 'post'
  385. # chrome/chromium only supports HTTP GET....
  386. if request.headers.get('User-Agent', '').lower().find('webkit') >= 0:
  387. method = 'get'
  388. ret = render('opensearch.xml',
  389. opensearch_method=method,
  390. host=get_base_url())
  391. resp = Response(response=ret,
  392. status=200,
  393. mimetype="application/xml")
  394. return resp
  395. @app.route('/favicon.ico')
  396. def favicon():
  397. return send_from_directory(os.path.join(app.root_path,
  398. 'static',
  399. get_current_theme_name(),
  400. 'img'),
  401. 'favicon.png',
  402. mimetype='image/vnd.microsoft.icon')
  403. def run():
  404. app.run(
  405. debug=settings['server']['debug'],
  406. use_debugger=settings['server']['debug'],
  407. port=settings['server']['port']
  408. )
  409. application = app
  410. if __name__ == "__main__":
  411. run()