weavejester/clucy

Clojure interface to Lucene

Clucy

Clucy is a Clojure interface to Lucene.

Installation

To install Clucy, add the following dependency to your project.clj file:

[clucy "0.4.0"]

Usage

To use Clucy, first require it:

(ns example
  (:require [clucy.core :as clucy]))

Then create an index. You can use (memory-index), which stores the search index in RAM, or (disk-index "/path/to/a-folder"), which stores the index in a folder on disk.

(def index (clucy/memory-index))

Next, add Clojure maps to the index:

(clucy/add index
   {:name "Bob", :job "Builder"}
   {:name "Donald", :job "Computer Scientist"})

You can remove maps just as easily:

(clucy/delete index
   {:name "Bob", :job "Builder"})

Once maps have been added, the index can be searched:

user=> (clucy/search index "bob" 10)
({:name "Bob", :job "Builder"})

user=> (clucy/search index "scientist" 10)
({:name "Donald", :job "Computer Scientist"})

You can search and remove all in one step. To remove all of the scientists...

(clucy/search-and-delete index "job:scientist")

Storing Fields

By default all fields in a map are stored and indexed. If you would like more fine-grained control over which fields are stored and index, add this to the meta-data for your map.

(with-meta {:name "Stever", :job "Writer", :phone "555-212-0202"}
  {:phone {:stored false}})

When the map above is saved to the index, the phone field will be available for searching but will not be part of map in the search results. This example is pretty contrived, this makes more sense when you are indexing something large (like the full text of a long article) and you don't want to pay the price of storing the entire text in the index.

Default Search Field

A field called "_content" that contains all of the map's values is stored in the index for each map (excluding fields with {:stored false} in the map's metadata). This provides a default field to run all searches against. Anytime you call the search function without providing a default search field "_content" is used.

This behavior can be disabled by binding content to false, you must then specify the default search field with every search invocation.

clucy's Issues

Is there any particular reason why max-results is not an optional parameter?

I was about to write a patch to use IndexReader.numDocs() [0], but it turns out that would either break backward compatibility or it be half-assed (just a minimal form of search without any options), so I am putting it up here for discussion (assuming I haven't missed something blindingly obvious as to why max results needs to be always stringently limited).

[0] http://lucene.apache.org/core/3_6_0/api/all/org/apache/lucene/index/IndexReader.html

Update Lucene libraries to version 4.7.2+

Clojars version 0.3.1 does not have the latest changes for Lucene 4.2

The following dependency in project.clj "[clucy "0.3.1"]" leads to getting the old code that does not compile/ work with Lucene 4.2.

For example there is outdated "IndexWriter$MaxFieldLength" and there is no IndexReader.

So I suggest either updating the Clojars or changing Clucy version in project.clj master branch, because it's misleading to have 2 different version here and on Clojars both versioned "0.3.1".

sorting?

How can you sort the search results?

Need a new maintainer

Neither James nor myself are actively using clucy anymore, so we should probably find someone who is who can help shepherd pull requests so they don't stagnate.

Relevant to #22, #21, #20, #19, and #16.