This is a very simple tool designed to crawl websites, producing data on what each page links to and what static assets they depend on.
The tool can be used as either a command line interface, to produce JSON formatted results, or via the web interface that can graph the dependencies between all internal pages and their assets from the target website.
It's worth noting that this tool works best on sites with few broken links, and that output can get rather large if the site has many pages.
Ensure your machine has node installed, along with node's package manager npm.
Once node & npm are installed, clone this repo and run
npm install from the project root. This will
install all the required dependencies, and run tests to ensure everything is configured correctly.
Once installed, you can run
webcrawler by executing
./bin/webcrawler from the project root. A usage
script will be printed if you run
webcrawler with no arguments.
crawl <target> <jsonFile>
crawl command can be used to recursively crawl the supplied target. If no JSON file target
is provided, then the result to be outputted to standard out. Otherwise the data is dumped to the
file provided as the
To access the web interface, run webcrawler with the argument
serve. This will boot a server on
port 3000 that is then accessible at
http://localhost:3000. From here, you can enter the target
URL and the results will be graphed onscreen.
Log levels are set via the environment variable
LOG_LEVEL. To run with debug output, try
LOG_LEVEL=debug ./bin/webcrawler <cmd>....
Much of the d3 code was adapted from the flare Hierarchical Edge Bundling demo, found here, though now wrapped in an angular directive.
Lawrence Jones - email@example.com