fact-base
2018-03-28
Simple implementation of fact-base data storage for Common Lisp
Upstream URL
Author
License
fact-base
Simple implementation of fact-base data storage for Common Lisp
Requirements
:fact-base
depends on alexandria
, anaphora
, local-time
, optima
and cl-fad
. All of them installable via quicklisp
.
Example Usage
; SLIME 2013-11-17
CL-USER> (ql:quickload :fact-base)
To load "fact-base":
Load 1 ASDF system:
fact-base
; Loading "fact-base"
(:FACT-BASE)
CL-USER> (defpackage :demo (:use :cl :fact-base))
#<PACKAGE "DEMO">
CL-USER> (in-package :demo)
#<PACKAGE "DEMO">
DEMO> (defparameter *db* (make-fact-base))
*DB*
DEMO> (insert! *db* (list 0 :message "This is a sample message"))
NIL
DEMO> (insert! *db* (list 1 :message "This is another one"))
NIL
DEMO> (insert! *db* (list 1 :author "Inaimathi"))
NIL
DEMO> (insert! *db* (list 2 :message "That second one was written by me. This one is a meta-message."))
NIL
DEMO> (insert! *db* (list 2 :type :meta))
NIL
DEMO> (current *db*)
((2 :TYPE :META)
(2 :MESSAGE "That second one was written by me. This one is a meta-message.")
(1 :AUTHOR "Inaimathi") (1 :MESSAGE "This is another one")
(0 :MESSAGE "This is a sample message"))
DEMO> (history *db*)
((@2014-03-20T09:47:19.622527-04:00 :INSERT (2 :TYPE :META))
(@2014-03-20T09:47:16.186383-04:00 :INSERT
(2 :MESSAGE
"That second one was written by me. This one is a meta-message."))
(@2014-03-20T09:47:08.834689-04:00 :INSERT (1 :AUTHOR "Inaimathi"))
(@2014-03-20T09:47:03.936936-04:00 :INSERT (1 :MESSAGE "This is another one"))
(@2014-03-20T09:46:04.771430-04:00 :INSERT
(0 :MESSAGE "This is a sample message")))
DEMO> (write! *db*)
"fb-1007"
Interlude: The file fb-1007
now contains
((5132 49564 771430000) :INSERT (0 :MESSAGE "This is a sample message"))
((5132 49623 936936000) :INSERT (1 :MESSAGE "This is another one"))
((5132 49628 834689000) :INSERT (1 :AUTHOR "Inaimathi"))
((5132 49636 186383000) :INSERT
(2 :MESSAGE "That second one was written by me. This one is a meta-message."))
((5132 49639 622527000) :INSERT (2 :TYPE :META))
we now resume your regularly scheduled demo.
DEMO> (multi-insert!
*db*
'((:message "Getting sick of specifying IDs manually, so I'll use multi-insert!")
(:user "Inaimathi")
(:mood "Curious")))
3
DEMO> (current *db*)
((3 :MOOD "Curious") (3 :USER "Inaimathi")
(3 :MESSAGE
"Getting sick of specifying IDs manually, so I'll use multi-insert!")
(2 :TYPE :META)
(2 :MESSAGE "That second one was written by me. This one is a meta-message.")
(1 :AUTHOR "Inaimathi") (1 :MESSAGE "This is another one")
(0 :MESSAGE "This is a sample message"))
DEMO> (delta *db*)
((@2014-03-20T09:51:13.740572-04:00 :INSERT (3 :MOOD "Curious"))
(@2014-03-20T09:51:13.740565-04:00 :INSERT (3 :USER "Inaimathi"))
(@2014-03-20T09:51:13.740544-04:00 :INSERT
(3 :MESSAGE
"Getting sick of specifying IDs manually, so I'll use multi-insert!")))
DEMO> (multi-insert!
*db*
'((:message "This next one will just append the delta to the file created last time.")
(:user "Inaimathi")
(:mood "Still curious")))
4
DEMO> (write-delta! *db*)
"fb-1007"
Interlude 2: That same file now contains...
((5132 49564 771430000) :INSERT (0 :MESSAGE "This is a sample message"))
((5132 49623 936936000) :INSERT (1 :MESSAGE "This is another one"))
((5132 49628 834689000) :INSERT (1 :AUTHOR "Inaimathi"))
((5132 49636 186383000) :INSERT
(2 :MESSAGE "That second one was written by me. This one is a meta-message."))
((5132 49639 622527000) :INSERT (2 :TYPE :META))
((5132 49873 740544000) :INSERT
(3 :MESSAGE
"Getting sick of specifying IDs manually, so I'll use multi-insert!"))
((5132 49873 740565000) :INSERT (3 :USER "Inaimathi"))
((5132 49873 740572000) :INSERT (3 :MOOD "Curious"))
((5132 49927 302099000) :INSERT
(4 :MESSAGE
"This next one will just append the delta to the file created last time."))
((5132 49927 302129000) :INSERT (4 :USER "Inaimathi"))
((5132 49927 302139000) :INSERT (4 :MOOD "Still curious"))
That file wasn't cleared, then re-written. The delta was just appended to the bottom. Note that both write!
and write-delta!
take an optional filename as a parameter, so it's possible to break a fact-base up across files or back it up in another location
DEMO> (multi-insert!
*db*
'((:message "Ok, lets do some lookups now")
(:user "Inaimathi")
(:mood "Thoughtful")))
5
DEMO> (lookup *db* :a 0)
((0 :MESSAGE "This is a sample message"))
T
DEMO> (lookup *db* :a 2)
((2 :TYPE :META)
(2 :MESSAGE "That second one was written by me. This one is a meta-message."))
T
DEMO> (lookup *db* :b :nope)
NIL
NIL
DEMO> (lookup *db* :b :user)
((5 :USER "Inaimathi") (4 :USER "Inaimathi") (3 :USER "Inaimathi"))
T
DEMO> (lookup *db* :a 3)
((3 :MOOD "Curious") (3 :USER "Inaimathi")
(3 :MESSAGE
"Getting sick of specifying IDs manually, so I'll use multi-insert!"))
T
DEMO> (lookup *db* :a 4)
((4 :MOOD "Still curious") (4 :USER "Inaimathi")
(4 :MESSAGE
"This next one will just append the delta to the file created last time."))
T
DEMO> (lookup *db* :a 5)
((5 :MOOD "Thoughtful") (5 :USER "Inaimathi")
(5 :MESSAGE "Ok, lets do some lookups now"))
T
DEMO> (lookup *db* :a 5 :b :user)
WARNING: No relevant index found, traversing...
((5 :USER "Inaimathi"))
DEMO> (multi-insert!
*db*
'((:message "The index system is still naive. It only has separate indices on :a, :b and :c, you see, but it doesn't save itself the time by at least starting with the smallest known index yet.")
(:user "Inaimathi")
(:mood "Even more thoughtful")))
6
DEMO> (multi-insert!
*db*
'((:message "I'm going to fix that. In fact...")
(:user "Inaimathi")
(:mood "Even more thoughtful")))
7
DEMO> (lookup *db* :c "Even more thoughtful")
((7 :MOOD "Even more thoughtful") (6 :MOOD "Even more thoughtful"))
T
DEMO> (insert! *db* (list 6 :important nil))
NIL
DEMO> (multi-insert!
*db*
'((:message "Yes, you can add metadata after the fact.")
(:user "Inaimathi")
(:mood "Oh, right...")))
8
DEMO>
Function documentation
:make-fact-base
:file-name
:next-id!
:lookup
:select
:matching?
:insert!
:multi-insert!
:delete!
:project!
:project
:write-delta!
:write!
:update!
:load!
Notes
-
No, we're not going to implement
:initial
. Compressing stuff on disk would be a better approach than cutting history when you need to branch, and it would let us defer decisions about internal fact-base projection formats (specifically as pertains to the next note) -
Define a structure other than a list to store history. -Specifically, you want the ability to push to its end (since that's the only way we'll ever be doing it)
See also
The following posts talk about some of the design decisions and parts of the code of :fact-base