Java: "Control" External Application

Question

Is it possible to programmatically start an application from Java and then send commands to it and receive the program's output?

I'm trying to realize this scenario:

I want to access a website that uses lots of javascript and special html + css features -> the website isn't properly displayed in swt.browser or any of the other of the available Browser Widgets. But the website can be displayed without any problems in firefox. So I want to run a hidden instance of firefox, load the website and get the data. (It would be nice if FF can be embedded in a JFrame or so..)

Has anybody got an idea how to realize this? Any help would really be appreciated!

EDIT: The website loads some Javascript that does some html magic and loads some pictures. When I only read the html from the website I see nothing more than some JavaScript calls. But when the website is loaded in a Browser, it displays some images overlayed with text. That's what I'm trying to show the user of my app.

Answer 1

To start Firefox from within the application, you could use:

Runtime runtime = Runtime.getRuntime();
try {
    String path = "/path/to/firefox";
    Process process = runtime.exec(path + " " + url);
} catch (IOException e) {
    // ...
}

To manipulate processes once they have started, one can often use process.getInputStream() and process.getOutputStream(), but that would not help you in the case of Firefox.

You should probably look into ways of solving your specific problem other than trying to interact directly between your application and a browser instance. Consider either moving the whole interface into a Java gui, or doing a web app from the ground up -- not half and half.

Answer 2

See this article - it will teach you how to start a process, read its output and write to its input stream.

However this solution may be not be the best for your problem. What kind of data do you need to get from the Web Page? Would it be better to read the html with an HTTP GET and then parse it with an Html parser?

Answer 3

If you have a text-mode browser available (like links2 on linux) you might want to see how well that can render the page. For example, the command "links -dump http://someurl.com" will format the page as text and exit immediately, resulting in output that might be easily parseable using the methods that Ray Myers and kgiannakakis suggest.

Answer 4

If the website is static, you could use a web scraper like Jericho to load the URL, parse the HTML and wander your way through the DOM to the info you need.

Answer 5

Although a similar feature to what you describe is planned for FireFox in the future, it is not available yet. The feature is dubbed TaskFox, and from the linked wiki, "its aim is to allow users to quickly access information and perform tasks that would normally take several steps to complete."

News of the upcoming TaskFox feature just broke today, in fact. Perhaps you should consider a career being a psychic instead of a programmer.