Elastic Search by Example Part I

Elastic search has been gaining ground as yet another nosql solution. In this series we will look at elastic search by building an application and get to know some of its features.
This post is still a work in progress and will evolve over time as elastic co releases newer versions of elastic search.

Why Elastic search?

Applications which need to perform free form text search are ideal to use elastic as there backed store, I am however not so comfortable thinking of using elastic where there are lots of updates happening. Applications where there is a lot of analytics is also well suited for elastic.

Free text search

Before we get into the details of elastic search ,we need to understand what is the primary use of elastic, elastic is ideal if your application involves searching of free form text, examples of such applications include, legal document processing, text mining application, ad hoc search application.



Elastic search is built on top of lucene search library which is a powerful robust search library, but it's limited to use as a library, elastic allows you to scale your search soultion.


For the purposes of this series we will try to build a classifieds application which has search on items

A classified application

Lets try to build a classified ads example by, here the primary use case is to search for ads by using keywords.

Architecture of the application

above diagram illustrates at very high level the 3 layers of the application, the data layer is the most important to us as that's where elastic search comes into picture.



The grunt work
To get started, the obvious first step is to download elastic search from here, as of this writing the version available is 2.1

the distribution is packaged as a zip file elasticsearch-2.1.0.zip
unzip this to any location, the bin folder contains the scripts to start elastic search, I am going to use elasticsearch.bat as I am using windows

elasticsearch-2.1.0\bin>elasticsearch.bat

[2015-12-02 19:19:48,083][INFO ][node ] [Richard Rider] version[2.1.0], pid[6028], build[72cd1f1/2015-11-18T22:40:03Z]
[2015-12-02 19:19:48,084][INFO ][node                     ] [Richard Rider] initializing ...
[2015-12-02 19:19:48,251][INFO ][plugins                  ] [Richard Rider] loaded [], sites []
[2015-12-02 19:19:48,611][INFO ][env                      ] [Richard Rider] using [1] data paths, mounts [[Windows7_OS (C:)]], net usable_space [93.7gb], net total_s
pace [287.1gb], spins? [unknown], types [NTFS]
[2015-12-02 19:20:01,453][INFO ][node                     ] [Richard Rider] initialized
[2015-12-02 19:20:01,454][INFO ][node                     ] [Richard Rider] starting ...
[2015-12-02 19:20:02,257][INFO ][transport                ] [Richard Rider] publish_address {127.0.0.1:9300}, bound_addresses {127.0.0.1:9300}, {[::1]:9300}
[2015-12-02 19:20:02,297][INFO ][discovery                ] [Richard Rider] elasticsearch/CRv8mY8fTv--0lunNkta4Q
[2015-12-02 19:20:06,394][INFO ][cluster.service          ] [Richard Rider] new_master {Richard Rider}{CRv8mY8fTv--0lunNkta4Q}{127.0.0.1}{127.0.0.1:9300}, reason: ze
n-disco-join(elected_as_master, [0] joins received)
[2015-12-02 19:20:07,192][INFO ][gateway                  ] [Richard Rider] recovered [0] indices into cluster_state
[2015-12-02 19:20:07,310][INFO ][http                     ] [Richard Rider] publish_address {127.0.0.1:9200}, bound_addresses {127.0.0.1:9200}, {[::1]:9200}
[2015-12-02 19:20:07,311][INFO ][node                     ] [Richard Rider] started

the above logs show elastic search starting up, the bit highlighted in green is the name of the node, this name can be configured (more on this later), in this case "Richard Rider" is something that is assigned by elastic.

once the service has started, go to this url in the browser, you will see the below json response

{
  "name" : "Richard Rider",
  "cluster_name" : "elasticsearch",
  "version" : {
    "number" : "2.1.0",
    "build_hash" : "72cd1f1a3eee09505e036106146dc1949dc5dc87",
    "build_timestamp" : "2015-11-18T22:40:03Z",
    "build_snapshot" : false,
    "lucene_version" : "5.3.1"
  },
  "tagline" : "You Know, for Search"
}

Installing plugins
Elastic search has a plugin mechanism that allows you to add functionality to elastic search, tow such plugins are kopf and head which allows you to manage and visualize the cluster.

to install the plugins execute the below command from the bin folder

elasticsearch-2.1.0\bin>plugin install lmenezes/elasticsearch-kopf/2.0

elasticsearch-2.1.0\bin>plugin install mobz/elasticsearch-head


from your browser navigate to the following URLS

kopf: http://localhost:9200/_plugin/kopf

head: http://localhost:9200/_plugin/head




with this we conclude the first part of this series, in subsequent parts we will build out the application and get ourselves familiarized with elastic search.