Sunil's Notes: December 2013

I wanted to set up ElasticSearch on Amazon EC2 and these are the steps that i followed to do that

Connect to your Amazon EC2 instance using SSH and execute following command to download ElasticSearch v 90.9 on my instance
```
wget https://download.elasticsearch.org/elasticsearch/elasticsearch/elasticsearch-0.90.9.tar.gz
```
It took less than a minute to download the elasticsearch binary on the instance
Next i executed following command to unzip the elasticsearch-0.90.9.tar.gz
```
tar -xf elasticsearch-0.90.9.tar.gz
```
After extracting the elasticsearch binaries start it by executing following command
```
elasticsearch-0.90.9/bin/elasticsearch –f
```
After starting the ElasticSearch instance i could use the ES REST API from the SSH console on the same instance
```
curl -XGET http://localhost:9200/_search
```
But since i want to access it from my local machine/outside the instance i had to open the 9200 port on that instances firewall that i could do by changing the security group and adding 9200 and 9300 ports to it
Then you can use the public DNS for your EC2 instance and query the REST instance from your local machine like this
```
curl -XGET http://<yourinstancename>.us-west-1.compute.amazonaws.com:9200/_search
```

Elastic Search has nice support for geo spatial search. I wanted to try that out. So i started by creating a index called testgeo and i did add addresses of few walmart locations in it. Then i used Elastic Search to figure out what store location is close to my location. I followed these steps to implement this sample

First i did create a new Index named testgeo like this, while creating index i marked location field as geo_type, this makes it possible to make geo distance queries.


curl -XPUT 'localhost:9200/testgeo' -d '{
    "mappings": {
        "walmart": {
          "properties": {
            "shopName": {
                "type": "string"
             },
             "address": {
                "properties": {
                   "city": {
                      "type": "string"
                   },
                   
                   "state": {
                      "type": "string"
                   },
                   "streetName": {
                      "type": "string"
                   },
                   "location": {
                      "type": "geo_point"
                      }
                   }
                }
             }
             
          }
    }
}'

Then i did add few locations to my index for each of the location i used the latitude, longitude for address of the location and stored it in my testgeo index


curl -XPOST 'localhost:9200/testgeo/walmart' -d '{
   "storeName": "Walmart Supercenter - Marina",
   "address": {
      "streetName": "150 Beach Rd",
      "cityName": "Marina",
      "state": "CA",
      "zipCode": 93933,
      "location": [
         -121.800565,
         36.69359
      ] 
   }
}'
            
curl -XPOST 'localhost:9200/testgeo/walmart' -d '{
   "storeName": "Walmart Mkt - Modesto",
   "address": {
      "streetName": "\"1421 Coffee Rd",
      "cityName": "Modesto",
      "state": "CA",
      "zipCode": 95355,
      "location": [
         -120.976622,
         37.664054
      ] 
   }
}'
            
curl -XPOST 'localhost:9200/testgeo/walmart' -d '{
   "storeName": "Walmart Supercenter - Patterson",
   "address": {
      "streetName": "\"1030 Sperry Ave",
      "cityName": "Patterson",
      "state": "CA",
      "zipCode": 95363,
      "location": [
         -121.142528,
         37.464052
      ] 
   }
}'

Once my index data is in place i could query the index with type equal to geo_distance, and i had to give a location from which i want to search all the stores in 100 KM of the location


curl -XPOST 'localhost:9200/testgeo/repair/_search?pretty=true'  -d '
{
    "query": {
        "filtered": {
           "query": {
                "match_all": {}
            },
           "filter": {
               "geo_distance": {
                  "distance": 100,
                  "distance_unit": "km",
                  "address.location": {
                     "lat": 37.53,
                     "lon": -121.97
                  }
               } 
           }
        }
    }
}'

Elastic Search returned list of stores matching the criteria.

Sunil's Notes

Installing ElasticSearch on Amazon EC2

Using Elastic Search to implement Locate US/ Geo Saptial Search