Jsoup的源码修改-支持Cookie等头信息的设置

Jsoup是一个开源的html解析工具，使用css子选择器的语法解析元素，比古老的HtmlParser好用很多。在最新版本中，不支持设置Cookie等头信息来访问url，这样一些需要设置Cookie的网页就无法抓取了。为方便使用，我稍微修改了源码。

版本1.2.3中，访问url的api为：

String url = "";
Document doc = Jsoup.parse(new URL(url), 3000);

不能设置Cookie等头信息。

下面我们进行源码的修改。先浏览源码：
Jsoup.parse的源码为：

   public static Document parse(URL url, int timeoutMillis) throws IOException {
        return DataUtil.load(url, timeoutMillis);
    }

DataUtil.load的源码为：

    static Document load(URL url, int timeoutMillis) throws IOException {
        String protocol = url.getProtocol();
        Validate.isTrue(protocol.equals("http") || protocol.equals("https"), "Only http & https protocols supported");

        HttpURLConnection conn = (HttpURLConnection) url.openConnection();
        conn.setInstanceFollowRedirects(true);
        conn.setConnectTimeout(timeoutMillis);
        conn.setReadTimeout(timeoutMillis);
        conn.connect();

很明显，调用HttpURLConnection.setRequestProperty来设置头信息即可。

我在DataUtil.java中重载了load函数，并抽取了一个公用方法，如下：

    static Document load(URL url, Map<String, String> requestMap, int timeoutMillis) throws IOException {
        String protocol = url.getProtocol();
        Validate.isTrue(protocol.equals("http") || protocol.equals("https"), "Only http & https protocols supported");

        HttpURLConnection conn = (HttpURLConnection) url.openConnection();
        conn.setInstanceFollowRedirects(true);
        conn.setConnectTimeout(timeoutMillis);
        conn.setReadTimeout(timeoutMillis);
        // set request prop
        for (String key : requestMap.keySet()) {
			conn.setRequestProperty(key, requestMap.get(key));
		}
        conn.connect();
        
        return load(url, conn);
    }

    static Document load(URL url, int timeoutMillis) throws IOException {
        String protocol = url.getProtocol();
        Validate.isTrue(protocol.equals("http") || protocol.equals("https"), "Only http & https protocols supported");

        HttpURLConnection conn = (HttpURLConnection) url.openConnection();
        conn.setInstanceFollowRedirects(true);
        conn.setConnectTimeout(timeoutMillis);
        conn.setReadTimeout(timeoutMillis);
        conn.connect();

        return load(url, conn);
    }
    
    private static Document load(URL url, HttpURLConnection conn) throws IOException {
        int res = conn.getResponseCode();
        if (res != HttpURLConnection.HTTP_OK)
            throw new IOException(res + " error loading URL " + url.toString());
        
        String contentType = conn.getContentType();
        if (contentType == null || !contentType.startsWith("text/"))
            throw new IOException(String.format("Unhandled content type \"%s\" on URL %s. Must be text/*", 
                    contentType, url.toString()));
        
        InputStream inStream = new BufferedInputStream(conn.getInputStream());
        String charSet = getCharsetFromContentType(contentType); // may be null, readInputStream deals with it

        Document doc = readInputStream(inStream, charSet, url.toExternalForm());
        inStream.close();
        return doc;
    }

我在Jsoup.java中重载了parse函数：

    public static Document parse(URL url, Map<String, String> requestMap, int timeoutMillis) throws IOException {
        return DataUtil.load(url, requestMap, timeoutMillis);
    }

完成以上修改了，编写ant脚本打jar包即可。

调用例子：

		String url = "";
		Map<String, String> requestMap = new HashMap<String, String>();
		requestMap.put("Cookie", "");
		Document doc = Jsoup.parse(new URL(url), requestMap, 3000);

这样就可以设置Cookie信息了，可解决需要登录验证的页面爬取问题。