High performance real-time search, indexes store in Redis for Rails application

#1

Post by huacnlee » Wed, Dec 14, 2011, 4:42 AM UTC

Features

Real-time search High performance Segment words search and prefix match search Support ActiveRecord and Mongoid Sort results by one field Homophone search, pinyin search Conditions support

High performance real-time search, indexes store in Redis for Rails application
github.com

Re: High performance real-time search, indexes store in Redis for Rails application

#2

Post by ldng » Wed, Dec 14, 2011, 8:49 AM UTC

Two questions pop in my mind. How does it compare to Solr ? What's wrong with Solr ?

Re: High performance real-time search, indexes store in Redis for Rails application

#3

Post by jerryji » Wed, Dec 14, 2011, 9:01 AM UTC

ldng wrote: post #2

Two questions pop in my mind. How does it compare to Solr ? What's wrong with Solr ?

After skimming the doc (in Chinese), here're my two answers, Sir:

* It offers way fewer functionalities than does Solr

* There is nothing wrong with Solr, just that someone wants to try something new that fits his/her needs

Re: High performance real-time search, indexes store in Redis for Rails application

#4

Post by Argorak » Wed, Dec 14, 2011, 9:37 AM UTC

By the way, picky does something similar and has a redis backend as well:

    http://florianhanke.com/picky/details.html

The website also explains the difference to full-text-indexers like SOLR and elasticsearch.

Re: High performance real-time search, indexes store in Redis for Rails application

#5

Post by alexpopescu » Wed, Dec 14, 2011, 6:26 PM UTC

I could see such solutions used in cases where:

1. you are already using Redis for some other part of the application

2. you don't want yet to have Lucene/Solr/ElasticSearch

3. you don't need to support too advanced searches

On the other hand while building such a library could be a good learning exercise, I think that before using it in an application you should look for other already existing libraries and try to collaborate with the original authors.

--

All things NoSQL: http://nosql.mypopescu.com