AQE: A Query Engine

This is an implementation of a knowledge base, hacked together in Python 3 (it won't work in Python 2 for reasons of modules) for now to quickly iterate on ideas.

There are a few key points:

• A KnowledgeBase contains facts.
• A fact is a tuple: (relationship, subject, object). For example, ('is', 'sky', 'blue').
• A KnowledgeBase has three core methods: ask, retract, and tell.
• The ask method queries the KnowledgeBase to ascertain whether a fact is true. Either the subject or the object may be None, in which case all satisifiable facts are returned.
• The retract method tells the KnowledgeBase that the fact is no longer true. If it's rainy, we might retract our fact about the sky being blue.
• The tell method tells the KnowledgeBase that the fact is now true. For example, if it's rainy (and we've retracted the previous 'sky is blue' fact), we might tell the KnowledgeBase that ('is', 'sky', 'grey').
• A KnowledgeBase can also perform substitutions.
• An action contains positive and negative preconditions, retractions, and updates. The positive condition list contains facts that must be true for a knowledge base, and the negative condition list contains facts that must be false. If these preconditions hold, the retractions are applied, followed by the updates.
• See test_actions.py for an example.

Limitations

• Singleton facts aren't supported; that is, there is no way to make a KnowledgeBase assert that there is only one relationship → subject mapping. For example, the KnowledgeBase will admit that ('is', 'shrödingers cat', 'alive') and ('is', 'schrödingers cat', 'dead') are both true simultaneously.

TODO

• Inference: given a list of actions, how to go from one state to another. The first step would be single-step, then integrating a search into the inference.
• Rewrite in C++?