How to use Scrapy with both Splash and Tor over Privoxy in Docker Compose

Question

I'm trying to run a Scrapy spider with two 'extensions':

1. Splash for rendering JavaScript,
2. Tor-Privoxy to provide anonymity.

As an example, I'm using the scraper of quotes.toscrape.com in https://github.com/scrapy-plugins/scrapy-splash/tree/master/example. Here is my directory structure:

.
├── docker-compose.yml
└── example
    ├── Dockerfile
    ├── scrapy.cfg
    └── scrashtest
        ├── __init__.py
        ├── settings.py
        └── spiders
            ├── __init__.py
            └── quotes.py

where the example directory is cloned from the scrapy-splash repository. I've added the following docker-compose.yml file:

version: '3'

services:
  scraper:
    build: ./example
    environment:
      - http_proxy=http://tor-privoxy:8118
    links:
      - tor-privoxy
      - splash

  tor-privoxy:
    image: rdsubhas/tor-privoxy-alpine

  splash:
    image: scrapinghub/splash

where in the settings.py file I've changed the SPLASH_URL:

# SPLASH_URL = 'http://127.0.0.1:8050/'
SPLASH_URL = 'http://splash:8050'

Because Splash is running not on the localhost, but in a separate linked container named splash. The Dockerfile for the scraper is

FROM python:alpine
RUN apk --update add libxml2-dev libxslt-dev libffi-dev gcc musl-dev libgcc openssl-dev curl bash
RUN pip install scrapy scrapy-splash
COPY . /scraper
WORKDIR /scraper
CMD ["scrapy", "crawl", "quotes"]

The problem is that when I run this using docker-compose build and docker-compose up, I get the following logs:

Starting examplecompose_tor-privoxy_1
Starting examplecompose_splash_1
Recreating examplecompose_scraper_1
Attaching to examplecompose_splash_1, examplecompose_tor-privoxy_1, examplecompose_scraper_1
splash_1       | 2017-07-11 16:10:13+0000 [-] Log opened.
splash_1       | 2017-07-11 16:10:13.794595 [-] Splash version: 3.0
tor-privoxy_1  | 2017-07-11 16:10:13.568 7f08e999eee8 Info: Privoxy version 3.0.23
tor-privoxy_1  | 2017-07-11 16:10:13.568 7f08e999eee8 Info: Program name: privoxy
tor-privoxy_1  | Jul 11 16:10:13.578 [notice] Tor v0.2.6.10 (git-58c51dc6087b0936) running on Linux with Libevent 2.0.22-stable, OpenSSL 1.0.2d and Zlib 1.2.8.
tor-privoxy_1  | Jul 11 16:10:13.578 [notice] Tor can't help you if you use it wrong! Learn how to be safe at https://www.torproject.org/download/download#warning
splash_1       | 2017-07-11 16:10:13.795925 [-] Qt 5.9.1, PyQt 5.9, WebKit 602.1, sip 4.19.3, Twisted 16.1.1, Lua 5.2
splash_1       | 2017-07-11 16:10:13.796204 [-] Python 3.5.2 (default, Nov 17 2016, 17:05:23) [GCC 5.4.0 20160609]
tor-privoxy_1  | Jul 11 16:10:13.578 [notice] Configuration file "/etc/tor/torrc" not present, using reasonable defaults.
tor-privoxy_1  | Jul 11 16:10:13.581 [notice] Opening Socks listener on 127.0.0.1:9050
splash_1       | 2017-07-11 16:10:13.796541 [-] Open files limit: 1048576
tor-privoxy_1  | Jul 11 16:10:13.000 [notice] Parsing GEOIP IPv4 file /usr/share/tor/geoip.
splash_1       | 2017-07-11 16:10:13.796706 [-] Can't bump open files limit
tor-privoxy_1  | Jul 11 16:10:13.000 [notice] Parsing GEOIP IPv6 file /usr/share/tor/geoip6.
splash_1       | 2017-07-11 16:10:13.903844 [-] Xvfb is started: ['Xvfb', ':1896918638', '-screen', '0', '1024x768x24', '-nolisten', 'tcp']
splash_1       | QStandardPaths: XDG_RUNTIME_DIR not set, defaulting to '/tmp/runtime-root'
tor-privoxy_1  | Jul 11 16:10:13.000 [warn] You are running Tor as root. You don't need to, and you probably shouldn't.
splash_1       | 2017-07-11 16:10:13.984515 [-] proxy profiles support is enabled, proxy profiles path: /etc/splash/proxy-profiles
tor-privoxy_1  | Jul 11 16:10:13.000 [notice] Bootstrapped 0%: Starting
splash_1       | 2017-07-11 16:10:14.041562 [-] verbosity=1
splash_1       | 2017-07-11 16:10:14.041732 [-] slots=50
tor-privoxy_1  | Jul 11 16:10:13.000 [notice] Bootstrapped 5%: Connecting to directory server
splash_1       | 2017-07-11 16:10:14.041806 [-] argument_cache_max_entries=500
tor-privoxy_1  | Jul 11 16:10:13.000 [notice] Bootstrapped 80%: Connecting to the Tor network
splash_1       | 2017-07-11 16:10:14.043083 [-] Web UI: enabled, Lua: enabled (sandbox: enabled)
splash_1       | 2017-07-11 16:10:14.044088 [-] Site starting on 8050
splash_1       | 2017-07-11 16:10:14.044240 [-] Starting factory <twisted.web.server.Site object at 0x7f73a4e4b3c8>
tor-privoxy_1  | Jul 11 16:10:14.000 [notice] Bootstrapped 85%: Finishing handshake with first hop
scraper_1      | 2017-07-11 16:10:15 [scrapy.utils.log] INFO: Scrapy 1.4.0 started (bot: scrashtest)
scraper_1      | 2017-07-11 16:10:15 [scrapy.utils.log] INFO: Overridden settings: {'BOT_NAME': 'scrashtest', 'DUPEFILTER_CLASS': 'scrapy_splash.SplashAwareDupeFilter', 'HTTPCACHE_STORAGE': 'scrapy_splash.SplashAwareFSCacheStorage', 'NEWSPIDER_MODULE': 'scrashtest.spiders', 'SPIDER_MODULES': ['scrashtest.spiders']}
scraper_1      | 2017-07-11 16:10:15 [scrapy.middleware] INFO: Enabled extensions:
scraper_1      | ['scrapy.extensions.corestats.CoreStats',
scraper_1      |  'scrapy.extensions.telnet.TelnetConsole',
scraper_1      |  'scrapy.extensions.memusage.MemoryUsage',
scraper_1      |  'scrapy.extensions.logstats.LogStats']
scraper_1      | 2017-07-11 16:10:15 [scrapy.middleware] INFO: Enabled downloader middlewares:
scraper_1      | ['scrapy.downloadermiddlewares.httpauth.HttpAuthMiddleware',
scraper_1      |  'scrapy.downloadermiddlewares.downloadtimeout.DownloadTimeoutMiddleware',
scraper_1      |  'scrapy.downloadermiddlewares.defaultheaders.DefaultHeadersMiddleware',
scraper_1      |  'scrapy.downloadermiddlewares.useragent.UserAgentMiddleware',
scraper_1      |  'scrapy.downloadermiddlewares.retry.RetryMiddleware',
scraper_1      |  'scrapy.downloadermiddlewares.redirect.MetaRefreshMiddleware',
scraper_1      |  'scrapy.downloadermiddlewares.redirect.RedirectMiddleware',
scraper_1      |  'scrapy.downloadermiddlewares.cookies.CookiesMiddleware',
scraper_1      |  'scrapy_splash.SplashCookiesMiddleware',
scraper_1      |  'scrapy_splash.SplashMiddleware',
scraper_1      |  'scrapy.downloadermiddlewares.httpproxy.HttpProxyMiddleware',
scraper_1      |  'scrapy.downloadermiddlewares.httpcompression.HttpCompressionMiddleware',
scraper_1      |  'scrapy.downloadermiddlewares.stats.DownloaderStats']
scraper_1      | 2017-07-11 16:10:15 [scrapy.middleware] INFO: Enabled spider middlewares:
scraper_1      | ['scrapy.spidermiddlewares.httperror.HttpErrorMiddleware',
scraper_1      |  'scrapy_splash.SplashDeduplicateArgsMiddleware',
scraper_1      |  'scrapy.spidermiddlewares.offsite.OffsiteMiddleware',
scraper_1      |  'scrapy.spidermiddlewares.referer.RefererMiddleware',
scraper_1      |  'scrapy.spidermiddlewares.urllength.UrlLengthMiddleware',
scraper_1      |  'scrapy.spidermiddlewares.depth.DepthMiddleware']
scraper_1      | 2017-07-11 16:10:15 [scrapy.middleware] INFO: Enabled item pipelines:
scraper_1      | []
scraper_1      | 2017-07-11 16:10:15 [scrapy.core.engine] INFO: Spider opened
scraper_1      | 2017-07-11 16:10:15 [scrapy.extensions.logstats] INFO: Crawled 0 pages (at 0 pages/min), scraped 0 items (at 0 items/min)
scraper_1      | 2017-07-11 16:10:15 [scrapy.extensions.telnet] DEBUG: Telnet console listening on 127.0.0.1:6023
tor-privoxy_1  | Jul 11 16:10:16.000 [notice] Bootstrapped 90%: Establishing a Tor circuit
tor-privoxy_1  | Jul 11 16:10:17.000 [notice] Tor has successfully opened a circuit. Looks like client functionality is working.
tor-privoxy_1  | Jul 11 16:10:17.000 [notice] Bootstrapped 100%: Done
tor-privoxy_1  | Jul 11 16:10:17.000 [warn] Received http status code 404 ("Not found") from server '216.218.222.10:443' while fetching "/tor/keys/fp/585769C78764D58426B8B52B6651A5A71137189A+80550987E1D626E3EBA5E5E75A458DE0626D088C".
scraper_1      | 2017-07-11 16:10:29 [scrapy.core.engine] DEBUG: Crawled (200) <GET http://quotes.toscrape.com/> (referer: None)
scraper_1      | 2017-07-11 16:10:29 [scrapy.spidermiddlewares.offsite] DEBUG: Filtered offsite request to 'www.goodreads.com': <GET https://www.goodreads.com/quotes>
scraper_1      | 2017-07-11 16:10:29 [scrapy.spidermiddlewares.offsite] DEBUG: Filtered offsite request to 'scrapinghub.com': <GET https://scrapinghub.com>
tor-privoxy_1  | Jul 11 16:10:44.000 [notice] Have tried resolving or connecting to address '[scrubbed]' at 3 different places. Giving up.
tor-privoxy_1  | Jul 11 16:10:44.000 [notice] Have tried resolving or connecting to address '[scrubbed]' at 3 different places. Giving up.
scraper_1      | 2017-07-11 16:10:44 [scrapy.downloadermiddlewares.retry] DEBUG: Retrying <GET http://quotes.toscrape.com/tag/adulthood/page/1/ via http://splash:8050/render.json> (failed 1 times): 500 Internal Server Error
scraper_1      | 2017-07-11 16:10:44 [scrapy.downloadermiddlewares.retry] DEBUG: Retrying <GET http://quotes.toscrape.com/tag/be-yourself/page/1/ via http://splash:8050/render.json> (failed 1 times): 500 Internal Server Error
tor-privoxy_1  | Jul 11 16:10:55.000 [notice] Have tried resolving or connecting to address '[scrubbed]' at 3 different places. Giving up.
tor-privoxy_1  | Jul 11 16:10:55.000 [notice] Have tried resolving or connecting to address '[scrubbed]' at 3 different places. Giving up.
scraper_1      | 2017-07-11 16:10:55 [scrapy.downloadermiddlewares.retry] DEBUG: Retrying <GET http://quotes.toscrape.com/tag/success/page/1/ via http://splash:8050/render.json> (failed 1 times): 500 Internal Server Error
scraper_1      | 2017-07-11 16:10:55 [scrapy.downloadermiddlewares.retry] DEBUG: Retrying <GET http://quotes.toscrape.com/tag/books/page/1/ via http://splash:8050/render.json> (failed 1 times): 500 Internal Server Error
tor-privoxy_1  | Jul 11 16:10:56.000 [notice] Have tried resolving or connecting to address '[scrubbed]' at 3 different places. Giving up.
scraper_1      | 2017-07-11 16:10:56 [scrapy.downloadermiddlewares.retry] DEBUG: Retrying <GET http://quotes.toscrape.com/ via http://splash:8050/render.json> (failed 1 times): 500 Internal Server Error
tor-privoxy_1  | Jul 11 16:10:57.000 [notice] Have tried resolving or connecting to address '[scrubbed]' at 3 different places. Giving up.
tor-privoxy_1  | Jul 11 16:10:57.000 [notice] Have tried resolving or connecting to address '[scrubbed]' at 3 different places. Giving up.
scraper_1      | 2017-07-11 16:10:57 [scrapy.downloadermiddlewares.retry] DEBUG: Retrying <GET http://quotes.toscrape.com/tag/classic/page/1/ via http://splash:8050/render.json> (failed 1 times): 500 Internal Server Error
scraper_1      | 2017-07-11 16:10:57 [scrapy.downloadermiddlewares.retry] DEBUG: Retrying <GET http://quotes.toscrape.com/tag/aliteracy/page/1/ via http://splash:8050/render.json> (failed 1 times): 500 Internal Server Error

where I've interrupted the process for brevity. It seems like the scraper and tor-privoxy services are alternately complaining about a 500 Internal Service Error and not being able to 'resolve or connect to address', respectively.

I'm struggling to figure out why the http_proxy and Splash don't 'work together'. Can anyone point me in the right direction?

Answer 1

Following the Aquarium template project (https://github.com/TeamHG-Memex/aquarium), I found that the trick is to make Splash use Tor, not the spider directly.

My adapted project has the following structure:

.
├── docker-compose.yml
├── example
│   ├── Dockerfile
│   ├── scrapy.cfg
│   └── scrashtest
│       ├── __init__.py
│       ├── settings.py
│       └── spiders
│           ├── __init__.py
│           └── quotes.py
└── splash
    └── proxy-profiles
        └── default.ini

and the docker-compose.yml is

version: '3'

services:
  scraper:
    build: ./example
    links:
      - splash

  tor-privoxy:
    image: rdsubhas/tor-privoxy-alpine

  splash:
    image: scrapinghub/splash
    volumes:
      - ./splash/proxy-profiles:/etc/splash/proxy-profiles:ro
    links:
      - tor-privoxy

where I've mounted the proxy-profiles directory as a volume into the splash container following http://splash.readthedocs.io/en/stable/api.html#proxy-profiles. The default.ini reads

[proxy]

host=tor-privoxy
port=8118

(I also noticed it is essential to call it default.ini).

With this setup, upon docker-compose build and docker-compose up the scraper runs successfully using Splash.