pydalog

Python implementation of a very basic, inefficient datalog

Example

Create Datalog Program

from datalog import Datalog

program = Datalog() \
        .add_fact('mother', 'sara', 'akbar') \
        .add_fact('mother', 'soghra', 'sara') \
        .add_fact('father', 'ahmad', 'akbar') \
        .add_rule(head=('parent', 'X', 'Y'), body=(('father', 'X', 'Y'),)) \
        .add_rule(head=('parent', 'X', 'Y'), body=(('mother', 'X', 'Y'),)) \
        .add_rule(head=('married', 'X', 'Y'), body=(('mother', 'X', 'Z'), ('father', 'Y', 'Z'))) \
        .add_rule(head=('married', 'Y', 'X'), body=(('married', 'X', 'Y'),)) \
        .add_rule(head=('descendant', 'X', 'Y'), body=(('parent', 'X', 'Y'),)) \
        .add_rule(head=('descendant', 'X', 'Y'), body=(('parent', 'X', 'Z'), ('parent', 'Z', 'Y')))

Query Program

Compile the program and fetch new rules using extend_db() function

program.extend_db()

Query facts with query(literal) function. If the literal is a Fact (arguments are all constant and not variable) returns a boolean indicating whether this fact is true or not.

If literal is not a fact and has some variables, returns the binding that will make a correct fact from the given literal

program.query('descendant', 'X', 'akbar') # returns [{X: 'sara'}, {X: 'ahmad'}, {X: 'soghra'}]
program.query('married', 'sara', 'ahmad') # returns True
program.query('married', 'sara', 'akbar') # returns False

Built-In Literals

some arithmetic operators are built in =, !=, <, <=, >, >=

=, != can compare anything. But other built-ins need float as their arguments.

usage:

from datalog.types import Eq, Lt, Gt, Lte, Neq

       program()\
       .add_rule(head=('child', 'X'), body=(('age', 'X', 'Y'), Lt('Y', 12))) \

look and run run.py for more details