donnyzhao/alexandra

Python toolkit for writing Amazon Echo skills as web services

alexandra

Minimal Python library to remove the tedious boilerplate-y parts of writing Alexa skills. Alexandra is tested against Python 2.7 and 3.5.1.

Alexandra can be used as part of an AWS lambda function or a self-hosted server. There's a builtin WSGI app if you're in to that kind of thing.

Check out the api documentation for more details on what alexandra can do.

import alexandra

app = alexandra.Application()
name_map = {}

@app.launch
def launch_handler():
    return alexandra.reprompt('What would you like to do?')

@app.intent('MyNameIs')
def set_name_intent(slots, session):
    name = slots['Name']
    name_map[session.user_id] = name

    return alexandra.respond("Okay, I won't forget you, %s" % name)

@app.intent('WhoAmI')
def get_name_intent(slots, session):
    name = name_map.get(session.user_id)

    if name:
        return alexandra.respond('You are %s, of course!' % name)

    return alexandra.reprompt("We haven't met yet! What's your name?")

if __name__ == '__main__':
    app.run('0.0.0.0', 8080, debug=True)

installing

Alexandra uses pyOpenSSL, which requires the libffi library to compile. Make sure that's installed first.

If you're on OS X, check out the special instructions for installing the OpenSSL library if you get errors during installation.

And then:

pip install alexandra

using alexandra with aws lambda

Getting an alexandra app running on lambda is much easier than running your own server, and is probably the right choice unless you need to access the local network or have some other complication that prevents you from using the service.

Here's an example:

app = alexandra.Application()

@app.intent('FooBar')
def foo_bar():
    ...

# Entry point to our lambda function.
def lambda_handler(event, context):
    return alexa.dispatch_request(event)

running with uwsgi

The alexandra.Application class has a run method, which is useful enough for testing purposes and simple projects, but for real deployments, you'll probably want to use something a little more robust, such as uWSGI.

Alexandra works with uwsgi in almost exactly the same way Flask does.

# skill_module.py

app = alexandra.Application()
wsgi_app = app.create_wsgi_app()

@app.intent('FooBar')
def foobar():
    ...

The above can be run with uwsgi as uwsgi -w skill_module:wsgi_app --http 0.0.0.0:5678

setting up a web server

Amazon requires a real SSL certificate for skills to be rolled out to other users, but fortunately for testing and personal projects self-signed certificates are acceptable.

To make it a bit easier to generate a self signed SSL certificate and nginx configuration, here's a Python 2.7/3.5.1 compatible standalone script to generate basic config for a standard unix setup.

After running the script, simply add a location block to the nginx config for any new Alexa skills being hosted on the same box.

For example, if there's an alexandra skill running on port 6789, you would add:

location /some_random_endpoint {
    proxy_pass http://localhost:6789;
}