I was going through example demos on creating Alexa skill sets and found that Alexa needs to use lambda function to give response to specific/similar type question. I was thinking about a skill with various simple question which doesn't require much of processing and, adding new set of question and answer should be as easy as writing down answer for questions in a notebook rather than modifying source code of the skill set. Thats when I got an idea about utilizing Api.AI to add new question and its possible answers.
What it does
It takes different types of question as input and replies by checking for possible answers in Api.AI.
How I built it
I used apiai python module to send intents to Api.AI. apiai module gets response from its server and is used as answer for the question asked to Alexa
Challenges I ran into
One of the biggest challenge I ran into was creating lambda function for the skill. Since the skill requires external python modules like apiai, nump, json, etc. and I had no previous experience with lambda function, I hand to spend lot of time to figure out how to use isolated environment for python with virtualenv to create such function.
Accomplishments that I'm proud of
Able to integrate Alexa with Api.AI. Since api.ai is one of its own kind similar to Alexa with its own Natural Language Processing (NLP), it took a good amount of time to plan how to integrate them together to make Alexa even more powerful.
What I learned
I learned how to use AWS Lambda function and make Alexa Skill sets
What's next for Jester
After the hackathon I have plans to improve the way Alexa is currently integrated with Api.ai and publish it on the store.