XHTML to PDF using flying-saucer how to cache css

Question

In my production pipeline I need to generate a few hundred PDF from HTML. For this scenario I first convert the HTML into XHTML. Than im passing the 'cleaned' XHTML and the uri to the renderer.

Since the *.css and imageFiles are equal for all the XHTML files I dont need to resolve them all the time I process a file. Im successfully using the following code for caching images. How can I cache .css files aswell? I want to avoid putting all files into my classpath.

ITextRenderer renderer = new ITextRenderer();

ResourceLoaderUserAgent callback = new ResourceLoaderUserAgent(renderer.getOutputDevice());
callback.setSharedContext(renderer.getSharedContext());

for (MyObject myObject : myObjectList) {

    OutputStream os = new FileOutputStream(tempFile);

    final DocumentBuilderFactory documentBuilderFactory = DocumentBuilderFactory.newInstance();
    documentBuilderFactory.setValidating(false);
    DocumentBuilder builder = documentBuilderFactory.newDocumentBuilder();
    org.w3c.dom.Document document = builder.parse(myObject.getLocalPath); // full path to .xhtml

    renderer.getSharedContext().setUserAgentCallback(callback);

    renderer.setDocument(document, myObject.getUri());
    renderer.layout();
    renderer.createPDF(os);

    os.flush();
    os.close();
}
    ...


private static class ResourceLoaderUserAgent extends ITextUserAgent
{
    public ResourceLoaderUserAgent(ITextOutputDevice outputDevice) {
        super(outputDevice);
    }

    protected InputStream resolveAndOpenStream(String uri) {
        InputStream is = super.resolveAndOpenStream(uri);
        System.out.println("IN resolveAndOpenStream() " + uri);

        return is;
    }
}

Answer 1

incase someone facing the same problem here is how I solved it. Since I wasnt able to cache the *.css files inside my CustomUserAgent I had to find another way. My solution uses Squid as http-proxy to cache all frequently used resources.

Inside my CustomUserAgent I only need to access this proxy by passing the proxy-object.

public class ResourceLoaderUserAgent extends ITextUserAgent {

public ResourceLoaderUserAgent(ITextOutputDevice outputDevice) {
    super(outputDevice);
}

protected InputStream resolveAndOpenStream(String uri) {    

    HttpURLConnection connection = null;
    URL proxyUrl = null;
    try {
        Proxy proxy = new Proxy(Proxy.Type.HTTP, new InetSocketAddress("localhost", 3128));
        proxyUrl = new URL(uri);
        connection = (HttpURLConnection) proxyUrl.openConnection(proxy);
        connection.connect();

    } catch (Exception e) {
        throw new RuntimeException(e);
    }

    java.io.InputStream is = null;
    try {
        is = connection.getInputStream();
    } catch (java.net.MalformedURLException e) {
        XRLog.exception("bad URL given: " + uri, e);
    } catch (java.io.FileNotFoundException e) {
        XRLog.exception("item at URI " + uri + " not found");
    } catch (java.io.IOException e) {
        XRLog.exception("IO problem for " + uri, e);
    }

    return is;
}
}

cached:

resolving css took 74 ms
resolving images took 225 ms

uncached:

resolving css took 15466 ms
resolving images took 11236 ms

as you can see, the differents between cached and uncached resources are significant