facile_search 0.0.1 → 0.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: df9c6537e798b558d1322a4f9b4708261e5a95b6
-  data.tar.gz: 87270a7bbd1d922517e48594e5a7c2b1b22d866c
+  metadata.gz: 924b90dc73809cebcb2268d303e1fcbbc7526b69
+  data.tar.gz: 9e20031e7689e879137eb4acda78816c3442c2c7
 SHA512:
-  metadata.gz: 90f3969afac89a9a114806a71d720a7b24376cab41e219ecd0cec50d7e3b054dbabd318e56cc399950fb95bcb92d2f3871389d4b9bbee928fa33f7e7bf9cd473
-  data.tar.gz: 06b25f0d288f36320cbca8fe019a7cd69d273559da08a2724a09b09ada2a580a1f8721a3c590baca9f4fead1ccd265b30d3c1f572f153b57c890d1543e5b0087
+  metadata.gz: 9bedca991f4cd059d387884e23a407c6f8f614026d0fcc75350d0798618b43bcc8bd458f7189911e1a6ad57cef04d97a807d59baa8fb4a19f85709085a076f4a
+  data.tar.gz: a02e9b6412073de167582196132ab2e22fc82d3e26ccdf1d0b4e0ddbca96dba09a3da169b0938254f4b616510fa85a34c169992aaae7a82ae3ff29719f9c535a

data/.travis.yml

@@ -1,4 +1,7 @@
 language: ruby
 rvm:
+  - 2.1.5
   - 2.2.1
+services:
+  - redis-server
 

data/README.md

@@ -1,5 +1,8 @@
 # FacileSearch
 
+[![Build Status](https://travis-ci.org/i2bskn/facile_search.svg?branch=master)](https://travis-ci.org/i2bskn/facile_search)
+[![Code Climate](https://codeclimate.com/github/i2bskn/facile_search/badges/gpa.svg)](https://codeclimate.com/github/i2bskn/facile_search)
+
 Simple search with inverted index.  
 (Index data is stored in the Redis.)
 
@@ -21,12 +24,19 @@ Or install it yourself as:
 
 ## Usage
 
+Create inverted index object.
+
+```ruby
+meta_data = FacileSearch::MetaData.new(namespace: "example_index", text_field: "text", id_field: "id")
+index = FacileSearch::InvertedIndex.new(meta_data)
+```
+
+Indexing and search.
+
 ```ruby
-meta = FacileSearch::MetaData.new(namespace: "example_index", text_field: "text")
-index = FacileSearch::InvertedIndex.new(meta)
 sample = Struct.new(:id, :text).new(1, "some text")
 index.indexing(sample) # => "OK"
-index.search(["text"]) # => [1]
+index.search("text") # => [1]
 ```
 
 ## Contributing

data/lib/facile_search/inverted_index.rb

@@ -1,5 +1,3 @@
-require "facile_search/tokenizer/ngram"
-
 module FacileSearch
   class InvertedIndex
     extend Forwardable
@@ -30,11 +28,22 @@ module FacileSearch
       end
     end
 
-    def search(queries)
-      queries.each_with_object([]) {|query, obj|
-        tokens = tokenizer.tokenize(query)
-        return [] if tokens.size.zero?
-        obj << redis.hmget(namespace, *tokens).map {|ids| deserialize(ids) }.inject(&:&)
+    def search(*queries)
+      queries.flatten.each_with_object([]) {|query, obj|
+        if tokenizer.tokenizable?(query)
+          tokens = tokenizer.tokenize(query)
+          return [] if tokens.size.zero?
+          obj << redis.hmget(namespace, *tokens).map {|value| deserialize(value) }.inject(&:&)
+        else
+          cursor = 0
+          ids = []
+          10000.times do # TODO: Adjust max number of loop
+            cursor, matches = redis.hscan(namespace, cursor, match: "*#{query}*")
+            ids << matches.map {|_, value| deserialize(value) }
+            break if cursor.to_i.zero?
+          end
+          obj << ids.uniq
+        end
       }.inject(&:&)
     end
 

data/lib/facile_search/tokenizer/ngram.rb

@@ -1,12 +1,16 @@
 module FacileSearch
-  module Tokenizer
-    class NGram
-      def initialize(n)
-        @n = n
+  class Tokenizer
+    class NGram < self
+      def initialize(n = nil)
+        @n = n || 2 # default is bigram
       end
 
-      def tokenize(strings)
-        strings.chars.each_cons(@n).map(&:join)
+      def tokenizable?(text)
+        text.size >= @n
+      end
+
+      def dividing(text)
+        text.chars.each_cons(@n).map(&:join)
       end
     end
   end

data/lib/facile_search/tokenizer/normalization_hook.rb

@@ -0,0 +1,10 @@
+module FacileSearch
+  class Tokenizer
+    module NormalizationHook
+      def self.apply(text)
+        NKF.nkf("-Wwm0Z1", text)
+      end
+    end
+  end
+end
+

data/lib/facile_search/tokenizer.rb

@@ -0,0 +1,31 @@
+require "facile_search/tokenizer/ngram"
+require "facile_search/tokenizer/normalization_hook"
+
+module FacileSearch
+  class Tokenizer
+    def tokenize(text)
+      dividing(preprocess(text))
+    end
+
+    def tokenizable?(text)
+      raise NotImplementedError
+    end
+
+    def dividing(text)
+      raise NotImplementedError
+    end
+
+    def preprocess(text)
+      hooks.inject(text) {|txt, hook| hook.apply(txt) }
+    end
+
+    def add_hook(hook)
+      hooks << hook
+    end
+
+    def hooks
+      @hooks ||= []
+    end
+  end
+end
+

data/lib/facile_search/utils.rb

@@ -0,0 +1,9 @@
+module FacileSearch
+  module Utils
+    def divide_queries(queries)
+      queries.split(/[[:blank:]]/)
+    end
+  end
+  extend Utils
+end
+

data/lib/facile_search/version.rb

@@ -1,4 +1,4 @@
 module FacileSearch
-  VERSION = "0.0.1"
+  VERSION = "0.0.2"
 end
 

data/lib/facile_search.rb

@@ -1,3 +1,4 @@
+require "nkf"
 require "forwardable"
 
 require "redis/objects"
@@ -5,5 +6,6 @@ require "oj"
 
 require "facile_search/version"
 require "facile_search/meta_data"
+require "facile_search/tokenizer"
 require "facile_search/inverted_index"
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: facile_search
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.0.2
 platform: ruby
 authors:
 - i2bskn
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-04-28 00:00:00.000000000 Z
+date: 2015-04-29 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: redis-objects
@@ -112,7 +112,10 @@ files:
 - lib/facile_search.rb
 - lib/facile_search/inverted_index.rb
 - lib/facile_search/meta_data.rb
+- lib/facile_search/tokenizer.rb
 - lib/facile_search/tokenizer/ngram.rb
+- lib/facile_search/tokenizer/normalization_hook.rb
+- lib/facile_search/utils.rb
 - lib/facile_search/version.rb
 homepage: https://github.com/i2bskn/facile_search
 licenses: