Project

redis-asm

0.01
No commit activity in last 3 years
No release in over 3 years
Fast ASM (Approximate String Matching) by calucuating edit distance within the collecitons such as ZSET, HASH, LIST, SET on Redis using Lua script. It provides you to search multi-byte characters correctly, because it recognizes lead-byte of UTF-8 strings.
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
 Dependencies

Development

~> 1.7
~> 10.0
>= 0

Runtime

>= 0
~> 3.0
 Project Readme

Redis::Asm

Build Status Coverage Status

Fast fuzzy string search on Redis using Lua. UTF-8 ready.

Description

Fast ASM (Approximate String Matching) by calculating edit distance within the collections such as ZSET, HASH, LIST, SET on Redis using Lua script.
Redis::Asm provides you to search multi-byte characters correctly, because it recognizes lead-byte of UTF-8 strings.

Prerequisites

This library requires a Redis server with Lua scripting support (EVAL and EVALSHA commands). This support was added in Redis 2.6.

Installation

Add this line to your application's Gemfile:

gem 'redis-asm'

And then execute:

$ bundle

Or install it yourself as:

$ gem install redis-asm

Usage

To initialize Redis::Asm:

require 'redis'
require 'redis-asm'

# Use Redis.current:
redis = Redis.current

# Initialize Redis with host and port:
redis = Redis.new(:host => REDIS_HOST, :port => REDIS_PORT)

asm = Redis::Asm.new(redis)

First, prepare test data:

data = %w(example samples abampere zzi 東京都 京都府)

# key names
keys = {}
types = ['set', 'zset', 'hash', 'list']
types.each{|t| keys[t] = "testdata:#{t}"}

# reset Redis
keys.values.each{|v| redis.del v }

# set data to Redis
redis.sadd         keys['set'],  data
redis.zadd         keys['zset'], data.map.with_index{|d, i| [i+1, d]} 
redis.mapped_hmset keys['hash'], ({}).tap{|h| data.each_with_index{|x,i| h[i+1] = x}}
data.each{|d| redis.rpush keys['list'], d }

To execute fuzzy search from Redis collections:

require 'json'
require 'yaml'

# asm.search(KEY, NEELDE, MAX_RESULTS=10)

# To search from SET
result = asm.search(keys['set'], 'example')
# To search from LIST
result = asm.search(keys['list'], 'example')

puts JSON.parse(result).to_yaml
# ---
# - haystack: example
#   match: 1
# - haystack: samples
#   match: 0.57142857142857
# - haystack: abampere
#   match: 0.5

# To search from HASH

# Redis::Asm matches HASH values
# each item has 'field' property

result = asm.search(HASH_KEY, 'example')
puts JSON.parse(result).to_yaml
# ---
# - haystack: example
#   field: '1'
#   match: 1
# - haystack: samples
#   field: '2'
#   match: 0.57142857142857
# - haystack: abampere
#   field: '3'
#   match: 0.5

# To search from ZSET
# each item has 'score' property

result = asm.search(ZSET_KEY, 'example')
puts JSON.parse(result).to_yaml
# ---
# - haystack: example
#   score: '1'
#   match: 1
# - haystack: samples
#   score: '2'
#   match: 0.57142857142857
# - haystack: abampere
#   score: '3'
#   match: 0.5

You can use UTF-8 multibyte chars:

result = asm.search(ZSET_KEY, '東京都')
puts JSON.parse(result).to_yaml
# ---
# - haystack: "東京都"
#   match: 1
# - haystack: "京都府"
#   match: 0.33333333333333

Performance

  • PC: MBP 2.6 GHz Intel Core i5 16GM DD3 RAM
  • OS: Mac OSX 10.9.5
  • Ruby 2.1.5p273 [x86_64-darwin13.0]
  • Redis server v=2.6.17 bits=64

You can try benchmarking Redis::Asm by running rake bench in console.
That's the result I've got on my machine.

krt@mbp% ruby bench/bench.rb
                             user     system      total        real
          a :   1000 wd  0.000000   0.000000   0.000000 (  0.016898)
          a :  10000 wd  0.000000   0.000000   0.000000 (  0.165706)
          a : 100000 wd  0.000000   0.000000   0.000000 (  1.468973)

        baz :   1000 wd  0.000000   0.000000   0.000000 (  0.014015)
        baz :  10000 wd  0.000000   0.000000   0.000000 (  0.091153)
        baz : 100000 wd  0.000000   0.000000   0.000000 (  0.651317)

    rifmino :   1000 wd  0.000000   0.000000   0.000000 (  0.017831)
    rifmino :  10000 wd  0.000000   0.000000   0.000000 (  0.108233)
    rifmino : 100000 wd  0.000000   0.000000   0.000000 (  0.772444)

mskelngesol :   1000 wd  0.000000   0.000000   0.000000 (  0.015920)
mskelngesol :  10000 wd  0.000000   0.000000   0.000000 (  0.092513)
mskelngesol : 100000 wd  0.000000   0.000000   0.000000 (  0.701796)

       元気です :   1000 wd  0.000000   0.000000   0.000000 (  0.002177)
       元気です :  10000 wd  0.000000   0.000000   0.000000 (  0.028857)
       元気です : 100000 wd  0.000000   0.000000   0.000000 (  0.279001)

NOTE: To be fair, it's suitable for less or eql than about 10,000 words, for Redis blocks it's requests while executing Lua script.

Acknowledgment

Contributing

  1. Fork it ( https://github.com/krt/redis-asm/fork )
  2. Create your feature branch (git checkout -b my-new-feature)
  3. Commit your changes (git commit -am 'Add some feature')
  4. Push to the branch (git push origin my-new-feature)
  5. Create a new Pull Request