limiter.py 5.6 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165
  1. # SPDX-License-Identifier: AGPL-3.0-or-later
  2. # lint: pylint
  3. # pyright: basic
  4. """Some bot protection / rate limitation
  5. To monitor rate limits and protect privacy the IP addresses are getting stored
  6. with a hash so the limiter plugin knows who to block. A redis database is
  7. needed to store the hash values.
  8. Enable the plugin in ``settings.yml``:
  9. - ``server.limiter: true``
  10. - ``redis.url: ...`` check the value, see :ref:`settings redis`
  11. """
  12. import re
  13. import string
  14. import random
  15. from flask import request
  16. from searx import redisdb
  17. from searx.plugins import logger
  18. from searx.redislib import incr_sliding_window, secret_hash
  19. name = "Request limiter"
  20. description = "Limit the number of request"
  21. default_on = False
  22. preference_section = 'service'
  23. logger = logger.getChild('limiter')
  24. block_user_agent = re.compile(
  25. r'('
  26. + r'unknown'
  27. + r'|[Cc][Uu][Rr][Ll]|[wW]get|Scrapy|splash|JavaFX|FeedFetcher|python-requests|Go-http-client|Java|Jakarta|okhttp'
  28. + r'|HttpClient|Jersey|Python|libwww-perl|Ruby|SynHttpClient|UniversalFeedParser|Googlebot|GoogleImageProxy'
  29. + r'|bingbot|Baiduspider|yacybot|YandexMobileBot|YandexBot|Yahoo! Slurp|MJ12bot|AhrefsBot|archive.org_bot|msnbot'
  30. + r'|MJ12bot|SeznamBot|linkdexbot|Netvibes|SMTBot|zgrab|James BOT|Sogou|Abonti|Pixray|Spinn3r|SemrushBot|Exabot'
  31. + r'|ZmEu|BLEXBot|bitlybot'
  32. # unmaintained Farside instances
  33. + r'|'
  34. + re.escape(r'Mozilla/5.0 (compatible; Farside/0.1.0; +https://farside.link)')
  35. + '|.*PetalBot.*'
  36. + r')'
  37. )
  38. PING_KEY = 'SearXNG_limiter.ping'
  39. TOKEN_KEY = 'SearXNG_limiter.token'
  40. def ping():
  41. redis_client = redisdb.client()
  42. user_agent = request.headers.get('User-Agent', 'unknown')
  43. x_forwarded_for = request.headers.get('X-Forwarded-For', '')
  44. ping_key = PING_KEY + user_agent + x_forwarded_for
  45. redis_client.set(secret_hash(ping_key), 1, ex=600)
  46. def get_token():
  47. redis_client = redisdb.client()
  48. if not redis_client:
  49. # This function is also called when limiter is inactive / no redis DB
  50. # (see render function in webapp.py)
  51. return '12345678'
  52. token = redis_client.get(TOKEN_KEY)
  53. if token:
  54. token = token.decode('UTF-8')
  55. else:
  56. token = ''.join(random.choice(string.ascii_lowercase + string.digits) for _ in range(8))
  57. redis_client.set(TOKEN_KEY, token, ex=600)
  58. return token
  59. def token_is_valid(token):
  60. valid = token == get_token()
  61. logger.debug("token is valid --> %s", valid)
  62. return valid
  63. def is_accepted_request() -> bool:
  64. # pylint: disable=too-many-return-statements
  65. redis_client = redisdb.client()
  66. user_agent = request.headers.get('User-Agent', 'unknown')
  67. x_forwarded_for = request.headers.get('X-Forwarded-For', '')
  68. if request.path == '/healthz':
  69. return True
  70. if block_user_agent.match(user_agent):
  71. logger.debug("BLOCK %s: %s --> detected User-Agent: %s" % (x_forwarded_for, request.path, user_agent))
  72. return False
  73. if request.path == '/search':
  74. c_burst_max = 2
  75. c_10min_max = 10
  76. ping_key = PING_KEY + user_agent + x_forwarded_for
  77. if redis_client.get(secret_hash(ping_key)):
  78. logger.debug('got a ping')
  79. c_burst_max = 15
  80. c_10min_max = 150
  81. else:
  82. logger.debug('missing a ping')
  83. c_burst = incr_sliding_window(redis_client, 'IP limit, burst' + x_forwarded_for, 20)
  84. c_10min = incr_sliding_window(redis_client, 'IP limit, 10 minutes' + x_forwarded_for, 600)
  85. if c_burst > c_burst_max or c_10min > c_10min_max:
  86. logger.debug("BLOCK %s: too many request", x_forwarded_for)
  87. return False
  88. if len(request.headers.get('Accept-Language', '').strip()) == '':
  89. logger.debug("BLOCK %s: missing Accept-Language", x_forwarded_for)
  90. return False
  91. if request.headers.get('Connection') == 'close':
  92. logger.debug("BLOCK %s: got Connection=close", x_forwarded_for)
  93. return False
  94. accept_encoding_list = [l.strip() for l in request.headers.get('Accept-Encoding', '').split(',')]
  95. if 'gzip' not in accept_encoding_list and 'deflate' not in accept_encoding_list:
  96. logger.debug("BLOCK %s: suspicious Accept-Encoding", x_forwarded_for)
  97. return False
  98. if 'text/html' not in request.accept_mimetypes:
  99. logger.debug("BLOCK %s: Accept-Encoding misses text/html", x_forwarded_for)
  100. return False
  101. if request.args.get('format', 'html') != 'html':
  102. c = incr_sliding_window(redis_client, 'API limit' + x_forwarded_for, 3600)
  103. if c > 4:
  104. logger.debug("BLOCK %s: API limit exceeded", x_forwarded_for)
  105. return False
  106. logger.debug(
  107. "OK %s: '%s'" % (x_forwarded_for, request.path)
  108. + " || form: %s" % request.form
  109. + " || Accept: %s" % request.headers.get('Accept', '')
  110. + " || Accept-Language: %s" % request.headers.get('Accept-Language', '')
  111. + " || Accept-Encoding: %s" % request.headers.get('Accept-Encoding', '')
  112. + " || Content-Type: %s" % request.headers.get('Content-Type', '')
  113. + " || Content-Length: %s" % request.headers.get('Content-Length', '')
  114. + " || Connection: %s" % request.headers.get('Connection', '')
  115. + " || User-Agent: %s" % user_agent
  116. )
  117. return True
  118. def pre_request():
  119. if not is_accepted_request():
  120. return 'Too Many Requests', 429
  121. return None
  122. def init(app, settings):
  123. if not settings['server']['limiter']:
  124. return False
  125. if not redisdb.client():
  126. logger.error("The limiter requires Redis") # pylint: disable=undefined-variable
  127. return False
  128. app.before_request(pre_request)
  129. return True