#Overview

Core engine component which does the following

• Analysis the user input

• Run through the ML algorithm

• Return the response

• Basic trainer model

• Training data set

• Test data set

• Memorize the outputs

• continuous and supervised learning via manual training on engine management process

• Simple REST API endpoint

Requirements

• Python 3.5
• pip
• virtualenvwrapper (optional)
• mongodb

Setup

git clone https://[email protected]/twbhackathon/skadoosh.git
cd skadoosh
[[ -n $VIRTUAL_ENV ]] && mkvirtualenv skadoosh -p `which python3.5`
[[ -z $VIRTUAL_ENV ]] && workon skadoosh
pip install -r engine/src/requirements.txt
python -m nltk.downloader all
touch engine/src/api/application.cfg

Enter the below 2 lines in application.cfg file (donot forget to modify the SECRET_KEY value :P)

SECRET_KEY   = "some secret key[edit]"
MONGO_DBNAME = "skadooshCoreDb"

Running the engine api

chmod +x run.sh
./run.sh

Let this terminal and open a new terminal tab/session navigate to repo directory and do a workon skadoosh

Note: Whenever you want to run the app in new terminal tab/session you need to execute once workon skadoosh

Integration Tests

chmod +x tests.sh
./tests.sh

Sample interactions

curl "http://localhost:4000/api/help" -d text="what's my account balance?"

Training Model

Soon...

FOR UI

Requirements

• node (brew install node)
• npm
• npm install --save pngquant-bin
• grunt cli(npm install -g grunt-cli.)
• bower(brew install bower)

go to directory skadoosh/skin/web/

grunt serve