Walker
Walker is a tool that performs a recursive analysis of a website, including all subdirectories and pages, to search for faulty links. It operates under the premise that the homepage of the site in question leads to other pages within the website, which subsequently lead to yet more pages.
It is worth noting that this process may take an indeterminate amount of time, as a website may contain an infinite number of nested links. A potential future update to the recursive function could be the addition of a depth parameter, which would restrict recursion to only a specified number of levels. I don't plan on implementing this any time soon due to CBSE board exams and college applications, though!
Also, walker is fast as fuck. After it acquires all the URLs in a website, it quickly(say, 0 seconds) performs status check for all of them concurrently. For example, it verified all the links(2026) on Next's website in 61 seconds:
Stats
Time to get all links: 61 seconds
Time to verify links: 0 seconds
Quirks
Client side rendering
Since the implementation of this tool works through fetching the HTML of the website in question, it would be impossible for it to retrieve HTML for pages that render on the client. Basically, only the HTML available when you view the page source is analyzed.
I could perhaps use something like headless chrome to get the HTML for these pages, but that only adds overhead to problem I was initially trying to solve.
Rate limits
Since I don't wait between each request, some websites might enforce their rate-limiting policies on walker, and hence cause it to error out for URLs which are working perfectly fine. In these cases, if the API is returning semantic error codes, walker will display something like 429 Too Many Requests.
walker has a timeout of 5 seconds between each request. If the URL does not return a response within 5 seconds, it will error out and show that the operation was timed out.
Head requests
To save on bytes, walker performs HEAD requests instead of GET requests. However, some websites might deny responding to this method, which can error out walker for them.
Examples
Examples of what the data would look like after the process is over can be found in the data directory.
Install
To install walker, you will need to have cargo configured in your environment. You can see how to do so here. Once you're done doing so, simply run cargo install link-walker, and let it install.
Usage
Using walker is easy. It has a dead-simple CLI interface that can be used to visually see the results of the analysis. Options for it are:
So, for example, doing:
...would result in:
But when used with the -r argument, it would result in something like:
)
To debug what URL walker is currently fetching, simply pass it a -d debug flag.
In situations where walker is unable to resolve some arbitrary URL, it will properly show the error while verifying all the links. Often, this error occurs because reqwest is unable to resolve its DNS. You can check if that is the case by running link-walker --url <URL> -s. If it does not return an error, there is probably something else going on with the URL that needs to be looked at.
Extra
Note that this is an experimental tool, and I only made this to learn more about Rust. I wrote a blog post detailing how I made it!