Rensa is a framework for storing and modifying facts, and realizing facts into natural text.

Features

• Concept extraction. Rensa can help you automatically glean some simple facts from natural text.
• Data access and control. You can add, remove, edit, or search for Rensa facts programmatically.
• Realization. Facts can be realized from their data structure format into natural language.
• Actor support. You can define actors (people) associated with a set of facts (their "brains"). Actors can persuade others to change their beliefs.
• Story encoding. You can build Story objects from a set of Rensa facts. Just like brains, stories can be edited and realized into natural text.

You can encode most valid JSON name-value pairs with Rensa, but there is a recommended specification to follow. Check out the wiki for more.

Getting Started

Install Python 2.7

At the moment, you will need Python 2.7 to run rensapy. You can check what version of Python you have by running:

python --version

If you do not have Python 2.7 installed, you can install it from the Python website.

Or, if you use Homebrew:

brew install python

If you're worried about multiple Python versions existing in harmony, you can use virtualenv.

Try the Demos

Once you have Python 2.7 installed, clone or download the repository. Next, take a look inside the demos folder. The brain_demo.py file is a good place to start. To run it, cd into the demos/ directory and type python brain_demo.py.

This file shows you how to load facts (assertions) from a JSON file into a Rensa brain. It also steps through how to change and search through a brain, as well as how to realize its contents.

Making a brain is this easy:

from Brain import *
myBrain = Brain()

If you'd rather load Rensa assertions from JSON, use:

myBrain = load_brain(['path/to/your_file.json'])

Now we can add a fact to our brain:

myBrain.add_assertion(
  {
    "l":["you"],
    "relation":"has_property",
    "r":["awesome"]
  }
)

Made a mistake? You can find and edit the contents of any brain. Let's change the tense of our first assertion:

myBrain.edit_assertion(1,"tense","future")

We can realize all the assertions in a brain like this:

for a in myBrain.get_assertions():
  print a.realize(myBrain,False)

Once we're done changing the brain, we can save it as a JSON file:

myBrain.save_brain('your/save/path/')

The remaining three demos demonstrate the basics of extracting simple assertions, making actors with brains, and building stories.

Applications

Rensa is not meant to be able to represent all forms of meaning, but it can encode enough to be useful for a wide range of applications. For example, it has been used to model dilemmas, adapt stories, and generate games.

License

Copyright (c) Sarah Harmon: this source code is free to use under the MIT License with author attribution. Refer to LICENSE.txt for details.

If you use this repository, please cite:

@phdthesis{harmon2017narrative,
  title={Narrative Encoding for Computational Reasoning and Adaptation},
  author={Harmon, Sarah M},
  year={2017},
  school={University of California, Santa Cruz}
}