Provides an implementation of McCarthy's `amb` operator with
binding forms and acceptance test operator.

Constraints solving functions that operate on a Constraint Description
which is a map describing a constraint description containing the mappings:
- :variables -> seq of LVars
- :formula -> list describing a predicate expression composed of a mix of
  the LVars in :variables and Clojure functions.

Most functions in rv work off of one or more of the following root
concepts:

- Entity: a hashmap with a :kb/id key mapped to a unique value and namespaced keys
- Table: a set of hashmaps or Entities
- Fact: a vector triple in the form [entity-id attribute value]
- Relation: a set of Facts pertaining to a particular Entity
- LVar: a logic variable that can bind to any value in its :range
- Ground: a concrete value
- Query: a set of Facts containing a mix of LVars and Grounds  
- Rules: a set of Facts describing synthetic relations
- Production: a pair of: antecedent query and consequent Facts
- KB: a set of Relations about many Entities and possibly containing Productions  
- Constraint Description: a set of LVars and a Formula describing the domain of their values
- Formula: a list describing a predicate expression of mixed LVars and clojure functions

A minimal implementation of Datalog.

I came across the Soundex algorithm when researching the retro KAMAS outlining application.
Soundex is a phonetic algorithm for indexing words by sound.

Provides internal unification functions.
DO NOT USE THIS NS.
There is no guarantee that it will remain stable or at all.

Common learning-related functions and protocols.

Version spaces are a binary classification, empirical learning algorithm.
The approach, as described in 'Version spaces: a candidate elimination approach
to rule learning' by Tom Mitchel (1977) takes training examples (currently
Tuples of a like-arity) and manages a 'version space'. A version space is a
map containing two 'hypotheses' :S and :G. The :G hypothesis corresponds to the
most general versions of the training data that are consistent with them and :S
is the most specific versions. When a version space is presented with a new
example it runs a 'candidate elimination' algorithm to modify the hypotheses :S
and :G accordingly. Examples can be marked as being 'positive' examples, meaning
that they are preferred instances. Anything not marked as 'positive' are taken as
negative examples. Once trained, a version space can be used to classify new
examples as 'positive' or 'negative'. If new examples are not covered by the
existing hypotheses then they are classified as 'ambiguous' instead.

The simplest possible production rules system that uses a set
of EAV tuples as its knowledge base.

Common search-related functions and protocols.

A* search implementation.