WarcMiddleware is a library that lets users save mirror backups of websites to their computer. It is an addon for the Python web crawler Scrapy that saves web server transactions (requests and responses) into a Web ARChive (WARC) file (ISO 28500). The transactions can then be played back or viewed, similar to using Archive.org's WayBackMachine. The WARC format is a standard method of saving these transactions.
There are two ways to use WarcMiddleware: (1) as a replacement ClientFactory or (2) as a DownloaderMiddleware. The former is recommended. As a ClientFactory, WarcMiddleware hooks into Scrapy's HTTP class and saves the raw requests and responses. The DownloaderMiddleware version configures itself to save the requests and responses that Scrapy sends it. The problem with this method is that Scrapy does not pass along the raw data and some of it is lost along the way.
WarcMiddleware requires:
For Windows, many of these packages can be downloaded from http://www.lfd.uci.edu/~gohlke/pythonlibs/.
The entire github repository serves as an example project which will download a website and save it as a WARC file. To try it, download the repository as a zip file and extract it. After installing the prerequisites listed above, run:
$ crawler.py --url http://www.eurogamer.net/
Scrapy will save the website into a file named out.warc.gz
The crawler supports downloading urls from a sitemap.xml file. As an example, this can be used to download all the posts from a Blogspot site. To get the sitemap.xml from a Blogspot site, append "sitemap.xml" to the url and save the file.
http://dogs.blogspot.com/sitemap.xml
Crawl the sitemap using:
$ crawler.py --sitemap sitemap.xml
Scrapy will save the website contents into a file name out.warc.gz
The crawler supports crawling an entire domain or many domains.
If crawling a single domain, use:
$ crawler.py --mirror --url http://example.com
Using --mirror is the same as using --domains example.com.
For multiple domains, the following will crawl anchor links on example.com that lead to example.com or to othersite.com. And from othersite.com, it will crawl anchor links that lead to either:
$ crawler.py --domains example.com,othersite.com --url http://example.com
Note that using --domains will override --mirror.
The --accept and --reject parameters affect whether or not each anchor link on a site is crawled. Each accepts a comma-separated list of regular expressions that should either be crawled or never crawled. This does not affect downloading external assets such as images or CSS files.
This example will not crawl anchor links that contain the string /search/?:
$ crawler.py --mirror --reject /search/\? --url http://example.com
A file containing a list of urls, one per line, can be supplied:
$ crawler.py --url-file index.txt
The crawler uses these urls as though they were supplied to the --url argument.
If no further arguments are given, the crawler will archive each url as well
as any external assets for that url, but will not crawl any anchor links. If
--domains is added, then it will crawl any anchor links matching the specified
domains. If --mirror is used instead of domains, then it will only crawl
anchor links that are on the domain of the first url in the file.
After creating a WARC file, the contents can be played back allowing the user to view the saved website. One way to view the saved site is to use warc-proxy. Warc-proxy creates a proxy that channels traffic from a web browser and responds to requests to view websites. Rather than sending the live website, warc-proxy sends back the saved website contents from the WARC file.
A Scrapy project is needed to use WarcMiddleware. To create one, from a command prompt run:
$ scrapy startproject crawltest
This will create a crawltest directory (dir) with another crawltest dir inside.
After this, choose one of the following methods to use WarcMiddleware.
Copy warcclientfactory.py and warcrecords.py next to scrapy.cfg in the outer crawltest dir. Also copy over the hanzo dir to the outer dir.
In the inner dir, open settings.py and add the following lines to the bottom:
DOWNLOADER_HTTPCLIENTFACTORY = 'warcclientfactory.WarcHTTPClientFactory'
This will enable the custom WarcMiddleware ClientFactory. Additionally, create a simple spider by copying the simplespider.py file into the spiders dir.
Finally, to start the spider, from a command prompt in the outer dir run:
$ scrapy crawl simplespider -a url=http://www.eurogamer.net/
This should output a WARC file named out.warc.gz
Copy warcmiddleware.py and warcrecords.py next to scrapy.cfg in the outer crawltest dir. Also copy over the hanzo dir to the outer dir.
In the inner dir, open settings.py and add the following lines to the bottom:
DOWNLOADER_MIDDLEWARES = {'warcmiddleware.WarcMiddleware': 820}
This will enable the WarcMiddleware. Additionally, create a simple spider by copying the simplespider.py file into the spiders dir.
Finally, to start the spider, from a command prompt in the outer dir run:
$ scrapy crawl simplespider -a url=http://www.eurogamer.net/
This should output a WARC file named out.warc.gz