fantoccini

A high-level API for programmatically interacting with web pages through WebDriver.

This crate uses the WebDriver protocol to drive a conforming (potentially headless) browser through relatively high-level operations such as "click this element", "submit this form", etc.

Most interactions are driven by using CSS selectors. With most WebDriver-compatible browser being fairly recent, the more expressive levels of the CSS standard are also supported, giving fairly powerful operators.

Forms are managed by first calling Client::form, and then using the methods on Form to manipulate the form's fields and eventually submitting it.

For low-level access to the page, Client::source can be used to fetch the full page HTML source code, and Client::raw_client_for to build a raw HTTP request for a particular URL.

Examples

These examples all assume that you have a WebDriver compatible process running on port 4444. A quick way to get one is to run geckodriver at the command line. The code also has partial support for the legacy WebDriver protocol used by chromedriver and ghostdriver.

The examples will be using panic! or unwrap generously when errors occur (see map_err) --- you should probably not do that in your code, and instead deal with errors when they occur. This is particularly true for methods that you expect might fail, such as lookups by CSS selector.

Let's start out clicking around on Wikipedia:

use fantoccini::{Client, Locator};
use futures::future::Future;
let c = Client::new("http://localhost:4444");

// let's set up the sequence of steps we want the browser to take
tokio::run(
    c
        .map_err(|e| {
            unimplemented!("failed to connect to WebDriver: {:?}", e)
        })
        .and_then(|c| {
            // first, go to the Wikipedia page for Foobar
            c.goto("https://en.wikipedia.org/wiki/Foobar")
        })
        .and_then(|mut c| c.current_url().map(move |url| (c, url)))
        .and_then(|(mut c, url)| {
            assert_eq!(url.as_ref(), "https://en.wikipedia.org/wiki/Foobar");
            // click "Foo (disambiguation)"
            c.find(Locator::Css(".mw-disambig"))
        })
        .and_then(|e| e.click())
        .and_then(|mut c| {
            // click "Foo Lake"
            c.find(Locator::LinkText("Foo Lake"))
        })
        .and_then(|e| e.click())
        .and_then(|mut c| c.current_url())
        .and_then(|url| {
            assert_eq!(url.as_ref(), "https://en.wikipedia.org/wiki/Foo_Lake");
            Ok(())
        })
        .map_err(|e| {
            panic!("a WebDriver command failed: {:?}", e);
        })
);

How did we get to the Foobar page in the first place? We did a search! Let's make the program do that for us instead:

// -- snip wrapper code --
// go to the Wikipedia frontpage this time
c.goto("https://www.wikipedia.org/")
    .and_then(|mut c| {
        // find the search form
        c.form(Locator::Css("#search-form"))
    })
    .and_then(|mut f| {
        // fill it out
        f.set_by_name("search", "foobar")
    })
    .and_then(|f| {
        // and submit it
        f.submit()
    })
    // we should now have ended up in the rigth place
    .and_then(|mut c| c.current_url())
    .and_then(|url| {
        assert_eq!(url.as_ref(), "https://en.wikipedia.org/wiki/Foobar");
        Ok(())
    })
// -- snip wrapper code --

What if we want to download a raw file? Fantoccini has you covered:

// -- snip wrapper code --
// go back to the frontpage
c.goto("https://www.wikipedia.org/")
    .and_then(|mut c| {
        // find the source for the Wikipedia globe
        c.find(Locator::Css("img.central-featured-logo"))
    })
    .and_then(|mut img| {
        img.attr("src")
            .map(move |src| (img, src.expect("image should have a src")))
    })
    .and_then(move |(img, src)| {
        // now build a raw HTTP client request (which also has all current cookies)
        img.client().raw_client_for(fantoccini::Method::GET, &src)
    })
    .and_then(|raw| {
        use futures::Stream;
        // we then read out the image bytes
        raw.into_body().map_err(fantoccini::error::CmdError::from).fold(
            Vec::new(),
            |mut pixels, chunk| {
                pixels.extend(&*chunk);
                futures::future::ok::<Vec<u8>, fantoccini::error::CmdError>(pixels)
            },
        )
    })
    .and_then(|pixels| {
        // and voilla, we now have the bytes for the Wikipedia logo!
        assert!(pixels.len() > 0);
        println!("Wikipedia logo is {}b", pixels.len());
        Ok(())
    })
// -- snip wrapper code --

For more examples, take a look at the examples/ directory.

CI testing

Big thanks to Sauce Labs for providing us with free cross-browser testing!