facts-db/cl-facts/README.md

Branch

Download

facts-db/cl-facts

1.1 The world is the totality of facts, not of things.

– Ludwig Wittgenstein, Tractatus Logico Philosophicus

cl-facts is a small in-memory graph database for Common Lisp. It features :


Requirements

You will need :


Usage

FACTS:ADD &rest SPECS

Adds facts (triples) to the database. Triples can be grouped by subject.

(facts:add ("Blade Runner" :is-a :movie
                           :director "Ridley Scott"
                           :actor "Harison Ford"
                           :actor "Rutger Hauer")
           ("Snow White" :is-a :movie
                         :director "William Cottrell"
                         :director "David Hand"))

or either

(facts:add (?movie :is-a :movie
                   :title "Blade Runner"
                   :director "Ridley Scott"
                   :actor "Harrison Ford"
                   :actor "Rutger Hauer"))

(facts:add (?movie :is-a :movie
                   :title "Snow White and the Seven Dwarfs"
                   :director "William Cottrell"
                   :director "David Hand"))

(with ((?s ?p ?o))
  (format t "~&~S ~S ~S~&" ?s ?p ?o))

=> NIL
FACTS.ANON::MOVIE-0000 :ACTOR "Harison Ford"
FACTS.ANON::MOVIE-0000 :ACTOR "Rutger Hauer"
FACTS.ANON::MOVIE-0000 :DIRECTOR "Ridley Scott"
FACTS.ANON::MOVIE-0000 :IS-A :MOVIE
FACTS.ANON::MOVIE-0000 :TITLE "Blade Runner"
FACTS.ANON::MOVIE-0001 :DIRECTOR "David Hand"
FACTS.ANON::MOVIE-0001 :DIRECTOR "William Cottrell"
FACTS.ANON::MOVIE-0001 :IS-A :MOVIE
FACTS.ANON::MOVIE-0001 :TITLE "Snow White"

The second version with ?movie will generate an anonymous symbol prefixed with movie-. It is considered a more clean and efficient way to abstract identifiers.


FACTS:RM &rest SPECS

(facts:rm (?movie :actor "Harison Ford"))

FACTS:WITH SPECS &body BODY

To follow Wittgenstein’s view of the world, all queries get turned into testing the presence or absence of triples (facts).

Variables are prefixed with a question mark symbol “?” and are wildcards, matching everything. Nested queries get their variables expanded, giving pattern matching abilities. For instance :

(with ((?s ?p ?o)
  (format t "~&~S ~S ~S~&" ?s ?p ?o))

=>
"Blade Runner" :ACTOR "Harison Ford"
"Blade Runner" :ACTOR "Rutger Hauer"
"Blade Runner" :DIRECTOR "Ridley Scott"
"Blade Runner" :IS-A :MOVIE
"Snow White" :DIRECTOR "David Hand"
"Snow White" :DIRECTOR "William Cottrell"
"Snow White" :IS-A :MOVIE

Multiple queries on the same subject can be grouped together easily :

(facts:with ((?movie :is-a :movie
                     :title ?title
                     :director ?director))
  (format t "~A directed ~A~%" ?director ?title))

=>
Ridley Scott directed Blade Runner
David Hand directed Snow White and the Seven Dwarfs
William Cottrell directed Snow White and the Seven Dwarfs

Negative facts specifications will remove matching facts from the results.

(with ((?s ?p ?o)
       (:not ?s :actor "Harison Ford"))
  (format t "~&~S ~S ~S~&" ?s ?p ?o))

=>
"Snow White" :DIRECTOR "David Hand"
"Snow White" :DIRECTOR "William Cottrell"
"Snow White" :IS-A :MOVIE

FACTS:*DB*

The current facts database.


FACTS:CLEAR-DB

Clears the database from every facts.


FACTS:SAVE-DB &key INTO (READABLY T)

Dump the database facts into filespec INTO.


FACTS:LOAD-DB SRC

Load the facts from SRC into *db*.


FACTS:WITH-TRANSACTION &body BODY

Enclose BODY database operations into a transaction.

A transaction ensures that all database operations will succeed or be reverted using their respective rollback functions.

Transactions can be nested safely.


TODO


Source

Download