SafariAlexa

SafariAlexa is an Alexa skill focussed on tourism, primarily tourism in India. The purpose of this skill is to help people to gather various information on different tourist places (especially in India as of now) from Alexa. This skill will allow Alexa to answer queries like what is the location of a particulat spot, what is the best time of the year to visit that spot, what are the special attractions of that place, etc. Instead of searching all these information from various sources on the Internet, users can simply ask their queries to Alexa and Alexa will answer them.

How to use the skill

The invocation phrase for this skill is go safari. So in order to use this skill preceed all your queries with the phrase 'ask go safari'. For example:

ask go safari to name some places to visit in maharashtra

ask go safari what is the location of manali

ask go safari to give information on manali

ask go safari what are the things I should do at manali

ask go safari what are the speacial attractions at manali

ask go safari to tell me the most suitable time to visit manali

ask go safari to name some places in manali

ask go safari to tell me some places similar to manali

ask go safari how should I reach manali

Technology Stack

Alexa NodeJS kit
NodeJS
AWS Lambda
Python
Flask
Flask-admin
Postgres
BeautifulSoup
Heroku

How the project works

The skill works at three stages, the Alexa voice user interface (VUI) stage, the client interacts with Alexa using VUI, that lambda function which runs the app logic for generating alexa response and finally the web app which provides the REST API from where Alexa fetches all its data.

When the user procvides a voice input, Alexa uses its VUI to handle the user input, and it finds out a particular intent which can serve the user query. Alexa, after determining the appropriate intent and parameter, sends a request to the lambda function. The lambda function uses the appropriate handler to serve the request. The handler actually makes a API call to the above mentioned REST API to get the required data, generates the response and sends it back to Alexa VUI.

The REST API

The REST API provides the neccessary data to server the user requests. The data is provided from a database which needs to be maintained by an admin.

Example of the REST API:

curl 'http://alexasafari.herokuapp.com/api/query_spot?spot=manali&query_type=info'
{
  "info": "Circled by towering peaks in the rich verdant valley of the Beas River, with mountain ventures waving from all directions, Manali is a year-round attraction. Travelers assemble here to hang out in the hippie villages around the main town; adventure seekers come for trekking, climbing, mountain biking, canyoning, paragliding, rafting and skiing; and much more.", 
  "spot_name": "manali", 
  "status": "FOUND"
}

Basic format

http://alexasafari.herokuapp.com/api/query_spot?spot=[name_of_the_spot]&query_type=[type_of_query]

http://alexasafari.herokuapp.com/api/query_state?state=[name_of_the_state]

where query_type can be any one of the following values

location
info
things_to_do
special_attraction
time_to_visit
near_by_places
similar_places
how_to_reach

Note: The skill is still in development statge and as of now it cannot answer queries for all the states.

The webapp maintains a database of tourist spots. The data is added by the admin. The server has two types of provision to server the requests it receives about spots and query type. If the information about the spot is available in its database, then it can return the data immidietely. Otherwise the server can employ scrapers to scrape the required information from the web. Presently the project contains one such scraper, the bttv_scraper which finds out the best time of the year to visit a particular place in case that particular place is not present in the data base. In future more such scrapers will be added.

The webapp provides a simple and easy to use admin interface to add and maintain the data base. Admin has to login using his credentials and an authorised admin can add other admins.

How to build the project

In order to build the skill along with the backend app and lambda functions, follow the below instructions

Creating the skill

Firstly, create a new ALexa skill by following the instructions present over here
Once the skill is created, open the tab for interaction model from the skill dash board, open the JSON edittor and paste the model from here
Next create a lambda function by following the instructions present over here
Do not forget to save and build your skill
In the edittor for the lambda function code, paste the content from here
Save your lambda function
Connect your lambda function to Alexa VUI

Building the backend app

Follow the below instructions to build the REST API locally

Enter into the directory called webapp
The API is build using python3, so you must have python3 installed
Execute pip install -r requirements.txt This will install all the dependencies for this app. If you encounter permission issues, please execute the above command with superuser access. For debian based systems like Ubuntu the command will be- sudo pip install -r requirements.txt You may use virtualenv for building this app if you do not want to install dependencies in your global space.

Installing the database

Execute the following to install Postgres on your system-

sudo apt-get update
sudo apt-get install postgresql postgresql-contrib

This will install Postgres on your system. Next you need to create a database where you will be creating tables for this project. By default Postgres creates a role (user) named after your system username (the username by which you are currently logged in) and a database names postgres. In order to create a new database and get familiar with Postgres follow this article- Setting up Postgres onn Ubuntu

After Postgres is up and running you will have to apply migrations for your app and create tables. For this execute the following from the project root.

python app/manage.py db init
python app/manage.py db migrate
python app/manage.py db upgrade

This will create tables and apply migrations.

Running the app

Finally run the app using the following:

python app/main.py

This will start the development server at 127.0.0.1 (localhost) and port 5000

Now visit the following URL in your browser:

htttp://127.0.0.1:5000

The admin interface can be accessed at /admin However for that you need to create an admin manually from your Postgres console. This will be improved in the future

Future goals

SafariAlexa is at the very early stages of development. Below are some future goals

Adding more data to the database
Adding more scrapers and data mining strategies for dynamically building a strong database
Improving interaction experience by adding more utterances

djmgit / SafariAlexa