TODO

Implement a system similar to wit.ai's for dataset generation (replacing the intent and entities dirs) so that it can become more flexible
Implement traits in the datasets and data processing pipeline
Simplify the entire API into a single class (Aurras) which contains all required functions
Implement Aurras' functions
 - Initialization
 - Create the dataset
    - Load the raw data
    - Create the dataset
    - Save the dataset
 - Load the dataset
 - Train a model
    - Create the model
    - Train the model
    - Save the model
 - Load a model
 - Respond to a prompt
 - Respond to a conversation
* - Parse speech
* - Generate speech
Implement Plugins
 - Using Google
 - Wolfrem Alpha
 - Music
 - Videos
 - Notion
Build a setup.py script
Build test scripts
Build a sample run script
Create documentation
Add logging to all core & integration scripts for better warning and error management
 - standardize
Create a response object to handel more complex responses
 - thing to say
 - questions to ask
 - thing to do (should interactions with integrations be managed by plugins or by Aurras?)