Outcome

Who Is Lingua?

A bot for your translation needs! With the help of Amazon Dot, ask Lingua to translate any word or sentence from English to another language. Lingua is a young bot, continuously learning new skills and languages to keep you on your feet, as you go through the world of language.

What Can She Do?

Provide Lingua with verbal input specifying a phrase and the language you want it translated to. She will respond verbally, as well as, send you an SMS with the translation.

Secondary functions include an option to revisit old translation and a quizzing game with questions based on previously translated phrases.

Tertiary Function could include a connection to a social API such as Facebook or Twitter. A user can share scores from the quiz or share progress in a language.

What's Lingua Like?

Lingua is a simple bot, constantly trying to keep you on your toes. When she's not busy helping you out, she's learning new languages and new ways to challenge you.

She is calm but doesn't do well with mediocrity. She is friendly but curt, encouraging but particular, understanding but witty. She can be quite the hypocrite, but who isn't?

Lingua In Action

What's Important?

The priority matrix below highlights the various tasks that needed to be programmed, arranged according to the importance and effort required to accomplish them.    

How Does She Work?

Lingua integrates three API's in it's functioning. It connects to the Microsoft Translate API for translations. The Alexa API provides the input and speech interaction while the Twilio API provides the SMS feedback. The app is written in Ruby Sinatra and PostgreSQL is used for the database. The app is deployed in the cloud using the Heroku platform service.

Functional Diagram

The diagram below illustrates the flow of possible interaction and functions.

Data Flow Diagram

The following diagram shows the various databases in the app uses and how the databases connect to each other and the Microsoft Translate API.

Looking Forward

Currently, Lingua can only take English inputs and provide answers for 16 languages which are all written in the Georgian script. In future, she hopes to learn languages in different scripts and provide translations from any language to another.

A daily word function can be introduced in the app which will send an SMS and verbally tell the user the translation of a new word every day.

Reflections

A few things I learned while working on this project were:

1. The various components and steps involved in writing and deploying an app.

2. The whole process is iterative and about troubleshooting. Google is your closest friend. Patience is key.

3. The most surprising aspect of this project was how easy it was to work with Alexa. The various functions and input parsing methods built in ensured working with Alexa was quick and easy to edit.

References

1. Microsoft Translate API:

API Location: https://datamarket.azure.com/dataset/bing/microsofttranslatorspeech

Documentation:http://docs.microsofttranslator.com/text-translate.html

Gem: https://github.com/relrod/bing_translator-gem

2. Alexa API:

Skill Creation: https://developer.amazon.com/edw/home.html#/skills/list
Documentation: https://developer.amazon.com/public/solutions/alexa/alexa-skills-kit/getting-started-guide
Gem: https://github.com/DanElbert/alexa_skills_ruby

3. Twilio:
Service Location: https://www.twilio.com
Documentation: https://www.twilio.com/docs/tutorials/walkthrough/server-notifications/ruby/sinatra
Gem: https://github.com/twilio/twilio-ruby


4. For everything else: https://github.com/daraghbyrne/onlineprototypes2016repo/tree/master/guides