Cupboard is an embedded database library for Clojure, intended to make writing Clojure objects to disk, and retrieving them, a relatively simple task. Cupboard is almost completely schemaless, supports unique and non-unique indices, querying by one or more indices, and ACID transactional semantics.

At its lowest level, Cupboard uses Berkeley DB JE, and contains a fairly complete Clojure wrapper for the JE API. This API is available in the cupboard.bdb.je package; some projects may find this code useful on its own. The cupboard.core package provides a higher-level abstraction.

Sample Code

This code makes some struct-maps representing books, saves them into a database, and updates them.

(ns examples.cupboard.simple
  (:use clojure.contrib.pprint)
  (:require [cupboard.core :as cb])
  (:use [cupboard.utils]))

(cb/open-cupboard! "/tmp/books")

(cb/defpersist book
  ((:isbn :index :unique)
   (:title :index :any)
   (:author :index :any)))

(cb/make-instance book ["0393039854", "The First Folio", "Shakespeare, William"])
(cb/make-instance book ["0393925870", "The Canterbury Tales", "Chaucer"])
(cb/make-instance book ["143851557X", "Troilus and Criseyde", "Chaucer"])
(cb/make-instance book ["019280619X", "Paradise Lost", "Milton, John"])

(println "this returns maps")
(pprint (cb/query (starts-with :isbn "039")))

(println "this returns struct-map values")
(pprint (cb/query (= :author "Chaucer") :struct book))

(println "this updates the :author slot")
(cb/query (= :author "Chaucer")
          :callback #(cb/passoc! % :author "Chaucer, Geoffrey"))

(println "this verifies that the update worked")
(pprint (cb/query (starts-with :author "Chaucer")))

(cb/close-cupboard!)

This example is available in examples/examples/cupboard/simple.clj. examples/examples/cupboard/gutenberg.clj uses a few more of Cupboard's features.

Dependencies

• Clojure (version 1.2.0 tested)
• clojure-contrib (version 1.2.0-compatible)
• Berkeley DB JE (version 4.0.92 tested)
• Joda Time (version 1.6.x tested)

Use Leiningen to download Cupboard's dependencies. Alternatively, Cupboard's build.xml contains an Ant task which downloads its dependencies, including Clojure and clojure-contrib.

Getting Started

1. Clone the Cupboard Git repository.
2. Run ant package. This pulls down all dependencies and puts them into the lib directory. Then it compiles Cupboard's source into Java class files and packages everything into a jar file.
3. Copy the Cupboard jar and its dependencies to wherever you normally copy your Clojure dependencies. Make sure to add them to your project's classpath.

Note: Full Leiningen support is forthcoming once Cupboard is more stable.

Documentation

Cupboard's documentation is still a work in progress. Please refer to the doc directory in the source distribution.

License

Cupboard is distributed under the MIT license, and so has few restrictions itself. Its Berkeley DB dependency, however, is licensed by Oracle — see Oracle's Berkeley DB Licensing Information page for details.