How to set up Python server side with javascript client side

Question

So there is already a Python program set up that runs on the console that I must build upon. I will be building a web GUI interface for the application using Javascript. How would I:

a. Go about handling the input/output of this Python program without touching the original code.

b. Go about sending console-line inputs to the Python program via Javascript calls. I've looked into raw HTTP requests/AJAX, but I'm not sure how exactly I would go about sending that as input to the Python program.

Answer 1

a. To handle input/output of the program: Pexpect. It's fairly easy to use, and reading some of the examples distributed with it should teach you enough to get the basics down.

b. Javascript interface:

Well, I use gevent and it's built-in WSGI server. (look up what a WSGI server (another) is). I should note that this program will keep a state, so you can manage your open sessions by returning a session ID to your javascript client and storing your pexpect session in a global variable or some other container so that you can complete the program's input and output across multiple independent AJAX requests. I leave that up to you, however, as that is not as simple.

All my example will do is put the POST request in some after clicking something of your choice. (it won't actually work because some of the variables are not set. Set them.)

Heres the relevant parts:

<!-- JavaScript -->
<script src="jquery.js"></script>
<script type="text/javascript">
function toPython(usrdata){
    $.ajax({
        url: "http://yoursite.com:8080",
        type: "POST",
        data: { information : "You have a very nice website, sir." , userdata : usrdata },
        dataType: "json",
        success: function(data) {
            <!-- do something here -->
            $('#somediv').html(data);
        }});
$("#someButton").bind('click', toPython(something));
</script>

Then the server:

# Python and Gevent
from gevent.pywsgi import WSGIServer
from gevent import monkey
monkey.patch_all() # makes many blocking calls asynchronous

def application(environ, start_response):
    if environ["REQUEST_METHOD"]!="POST": # your JS uses post, so if it isn't post, it isn't you
        start_response("403 Forbidden", [("Content-Type", "text/html; charset=utf-8")])
        return "403 Forbidden"
    start_response("200 OK", [("Content-Type", "text/html; charset=utf-8")])
    r = environ["wsgi.input"].read() # get the post data
    return r

address = "youraddresshere", 8080
server = WSGIServer(address, application)
server.backlog = 256
server.serve_forever()

If your program is Object Oriented, it'd be fairly easy to integrate this. EDIT: Doesn't need to be object oriented. and I have now included some Pexpect code

global d
d = someClass()
def application(environ, start_response):
    # get the instruction
    password = somethingfromwsgi # read the tutorials on WSGI to get the post stuff
    # figure out WHAT to do
    global d
    success = d.doSomething()
    # or success = funccall()
    prog = pexpect.spawn('python someprogram.py')
    prog.expect("Password: ")
    prog.sendline(password)
    i = prog.expect(["OK","not OK", "error"])
    if i==0:
        start_response("200 OK", [("Content-Type", "text/html; charset=utf-8")])
        return "Success"
    elif i==1:
        start_response("500 Internal Server Error", [("Content-Type", "text/html; charset=utf-8")])
        return "Failure"
    elif i==2:
        start_response("500 Internal Server Error", [("Content-Type", "text/html; charset=utf-8")])
        return "Error"

Another option I suggest is Nginx + uWSGI. If you would prefer that, I can give you some examples of that as well. It gives you the benefit of incorporating the webserver into the setup.