Printing a Numbered List of URLs

numbered-url-list.rs

/// This example demonstrates how print a numbered list of URLs from the Ubuntu Kernel PPA,
/// which can be used to automate the installation of Ubuntu kernels.

extern crate hyper;
use hyper::Client;
use hyper::header::Connection;

extern crate select;
use select::document::Document;
use select::predicate::{Name};

use std::io::Read;

fn main() {
    let input_url = "http://kernel.ubuntu.com/~kernel-ppa/mainline";
    for url in parse_document(input_url, get_page(input_url).as_str()) {
        println!(" {}. {}", url.id, url.url);
    }
}

fn get_page(input_url: &str) -> String {
    let client = Client::new();
    let mut response = client.get(input_url)
        .header(Connection::close()).send().unwrap();
    let mut body = String::new();
    response.read_to_string(&mut body).unwrap();
    body
}

struct URL {
    id:  u16,
    url: String,
}

fn parse_document(input_url: &str, document: &str) -> Vec<URL> {
    let mut id = 1;
    let mut output: Vec<URL> = Vec::new();
    for node in Document::from(document).find(Name("a")).iter() {
        if let Some(href) = node.attr("href") {
            if href.starts_with('/') {
                output.push(URL{id: id, url: String::from(input_url) + href});
            } else if href.starts_with('#') {
                output.push(URL{id: id, url: String::from(input_url) + "/"});
            } else if href.starts_with('j') {
                continue
            } else if !href.starts_with("http") {
                output.push(URL{id: id, url: String::from(input_url) + "/" + href});
            } else {
                output.push(URL{id: id, url: String::from(href)});
            }
        }
        id += 1;
    }
    output
}