scrapy-rotating-proxies
This package provides a Scrapy middleware to use rotating proxies, check that they are alive and adjust crawling speed.
License is MIT.
Installation
pip install rentier-scrapy-proxy-rotator
Usage
Add ROTATING_PROXY_LIST
option with a list of proxies to settings.py:
ROTATING_PROXY_LIST = [ 'proxy1.com:8000', 'proxy2.com:8031', # ... ]
As an alternative, you can specify a ROTATING_PROXY_LIST_PATH
options
with a path to a file with proxies, one per line:
ROTATING_PROXY_LIST_PATH = '/my/path/proxies.txt'
Both are used if provided.
Then add rotating_proxies middlewares to your DOWNLOADER_MIDDLEWARES:
DOWNLOADER_MIDDLEWARES = { # ... 'rentier_rotating_proxies.middlewares.RotatingProxyMiddleware': 610, 'rentier_rotating_proxies.middlewares.BanDetectionMiddleware': 620, # ... }
After this all requests will be proxied using one of the proxies from
the ROTATING_PROXY_LIST
/ ROTATING_PROXY_LIST_PATH
.
Requests with "proxy" set in their meta are not handled by
scrapy-rotating-proxies. To disable proxying for a request set
request.meta['proxy'] = None
; to set proxy explicitly use
request.meta['proxy'] = "<my-proxy-address>"
.
Concurrency
By default, all default Scrapy concurrency options (DOWNLOAD_DELAY
,
AUTHTHROTTLE_...
, CONCURRENT_REQUESTS_PER_DOMAIN
, etc) become
per-proxy for proxied requests when RotatingProxyMiddleware is enabled.
For example, if you set CONCURRENT_REQUESTS_PER_DOMAIN=2
then
spider will be making at most 2 concurrent connections to each proxy,
regardless of request url domain.
Customization
rentier-scrapy-proxy-rotator
keeps track of working and non-working proxies,
and re-checks non-working from time to time.
Detection of a non-working proxy is site-specific.
By default, scrapy-rotating-proxies
uses a simple heuristic:
if a response status code is not 200, response body is empty or if
there was an exception then proxy is considered dead.
You can override ban detection method by passing a path to
a custom BanDectionPolicy in ROTATING_PROXY_BAN_POLICY
option, e.g.:
# settings.py ROTATING_PROXY_BAN_POLICY = 'myproject.policy.MyBanPolicy'
The policy must be a class with response_is_ban
and exception_is_ban
methods. These methods can return True
(ban detected), False (not a ban) or None (unknown). It can be convenient
to subclass and modify default BanDetectionPolicy:
# myproject/policy.py from rotating_proxies.policy import BanDetectionPolicy class MyPolicy(BanDetectionPolicy): def response_is_ban(self, request, response): # use default rules, but also consider HTTP 200 responses # a ban if there is 'captcha' word in response body. ban = super(MyPolicy, self).response_is_ban(request, response) ban = ban or b'captcha' in response.body return ban def exception_is_ban(self, request, exception): # override method completely: don't take exceptions in account return None
Instead of creating a policy you can also implement response_is_ban
and exception_is_ban
methods as spider methods, for example:
class MySpider(scrapy.Spider): # ... def response_is_ban(self, request, response): return b'banned' in response.body def exception_is_ban(self, request, exception): return None
It is important to have these rules correct because action for a failed request and a bad proxy should be different: if it is a proxy to blame it makes sense to retry the request with a different proxy.
Non-working proxies could become alive again after some time.
scrapy-rotating-proxies
uses a randomized exponential backoff for these
checks - first check happens soon, if it still fails then next check is
delayed further, etc. Use ROTATING_PROXY_BACKOFF_BASE
to adjust the
initial delay (by default it is random, from 0 to 5 minutes). The randomized
exponential backoff is capped by ROTATING_PROXY_BACKOFF_CAP
.
Settings
ROTATING_PROXY_LIST
- a list of proxies to choose from;ROTATING_PROXY_LIST_PATH
- path to a file with a list of proxies;ROTATING_PROXY_LOGSTATS_INTERVAL
- stats logging interval in seconds, 30 by default;ROTATING_PROXY_CLOSE_SPIDER
- When True, spider is stopped if there are no alive proxies. If False (default), then when there is no alive proxies all dead proxies are re-checked.ROTATING_PROXY_PAGE_RETRY_TIMES
- a number of times to retry downloading a page using a different proxy. After this amount of retries failure is considered a page failure, not a proxy failure. Think of it this way: every improperly detected ban cost youROTATING_PROXY_PAGE_RETRY_TIMES
alive proxies. Default: 5.It is possible to change this option per-request using
max_proxies_to_try
request.meta key - for example, you can use a higher value for certain pages if you're sure they should work.ROTATING_PROXY_BACKOFF_BASE
- base backoff time, in seconds. Default is 300 (i.e. 5 min).ROTATING_PROXY_BACKOFF_CAP
- backoff time cap, in seconds. Default is 3600 (i.e. 60 min).ROTATING_PROXY_BAN_POLICY
- path to a ban detection policy. Default is'rotating_proxies.policy.BanDetectionPolicy'
.
FAQ
Q: Where to get proxy lists? How to write and maintain ban rules?
A: It is up to you to find proxies and maintain proper ban rules
for web sites; scrapy-rotating-proxies
doesn't have anything built-in.
There are commercial proxy services like https://crawlera.com/ which can
integrate with Scrapy (see https://github.com/scrapy-plugins/scrapy-crawlera)
and take care of all these details.