GroIMP wiki

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
{{howhard>3}}
====== Getting started the HTTP-server ======

GroIMP comes with an basic HTTP server that allows users to open projects from a given path.
The server can also manage basic additional input and the return of values to the http client.

In the following a brief start for this is given.


===== Starting the Server and Running a first Model =====

To start the server in GroIMP go on the main menu on Net/Open Http Server and select a port in the upcoming popup. This port is not specifically important, lets for now just use the default port suggested:"58080" (if you have another service running on this port just choose an other).


If we now go on [[http://localhost:58080]] GroIMP gives us a list of all installed plugins and the available http commands.

The command we are interested in is the open command, which is called with:

http://localhost:58080/open?<pathToyourGSZProject>

This works with every project and just opens the project, but the definition of the path follows some semi-intuitive rules:
  - you don't use a key: different to most URL based parameters you do write something like ?path=my/fancy/project.gsz. You just write ?my/fancy/project.gsz
  - you cant use absolute paths. all paths start form the Root directory defined in the GroIMP preferences under HTTP server/ Open Project.


==== Get a return value ====


The returned values for the client needs to be defined in the project it self. Therefore we need a custom startup function which is called very time the project is started.

Ours will now try to get the HttpResponse object for this workbench and if that exists tell GroIMP to start the run function with the response as a parameter.   

<code java>

import de.grogra.imp.net.*;

protected void startup()
{
	super.startup();
	HttpResponse resp = HttpResponse.get(workbench());
	if(resp!=null){
		runLater(resp);
	}
}

</code>


The Run function can than be defined as you wish and would normally include the execution of your model similar to a headless execution. In our case we skip this part and directly return our message to the client and then close the workbench:

<code java>

protected void run(Object info)
{
	HttpResponse resp = (HttpResponse)info; 
	resp.setContent("text/text","UTF-8","hello");
	resp.send(true);
	closeWorkbench();
}

</code>


If you save both of this files in a project and open it through the server it will return "hello" in our web browser. 

Additionally it is possible to access the original request from the project, which allows us to receive additional parameters given by he client.

For instance the following allows us to add ''&name=tim'' to the request and the project will greed me by name. 

<code java>
import de.grogra.imp.net.*;

protected void startup()
{
	super.startup();
	HttpResponse resp = HttpResponse.get(workbench());
	if(resp!=null){
		runLater(resp);
	}
}

protected void run(Object info)
{
	HttpResponse resp = (HttpResponse)info; 
	StringBuffer buf = new StringBuffer();
	buf.append("Hi ");
	buf.append(resp.getRequest().getQuery().split("&")[1].split("=")[1]);
	buf.append(", whats up ");
	resp.setContent("text/text","UTF-8",buf.toString());
	resp.send(true);
	closeWorkbench();
}
</code>
===== Starting the server headless =====

GroIMP allows users to define which of its are suppose to be executed as an commandline parameter. 
This allows us to start the http server headless on a port of our linking: 

<code>
java -jar core.jar --headless -- -cmd "/http/server=58080"
</code>