[SOLVED] Index Neo4j to work with Elasticsearch

Index Neo4j to work with Elasticsearch

I am trying to install Graph-Aided Search to integrate Neo4j with ElasticSearch (2.3.1) as shown here. But I'm struggling to define the indexes with curl.

When I try any of the curl commands like this one:

curl -XPUT http://localhost:9200/Person/_settings?index.gas.neo4j.user=neo4j

I get this error message:

{"error":{"root_cause":[{"type":"index_not_found_exception","reason":"no such index","resource.type":"index_or_alias","resource.id":"Person","index":"Person"}],"type":"index_not_found_exception","reason":"no such index","resource.type":"index_or_alias","resource.id":"Person","index":"Person"},"status":404}

Let's take for example the Movie database (Movie, Person). How can I define and configure Neo4j to index Movie and Person so I can call them with Elasticsearch?

UPDATE

When doing for example:

CURL XGET http://localhost:9200/person/_search?pretty=true

I get this result:

{
  "took" : 1,
  "timed_out" : false,
  "_shards" : {
    "total" : 5,
    "successful" : 5,
    "failed" : 0
  },
  "hits" : {
    "total" : 0,
    "max_score" : null,
    "hits" : [ ]
  }
}

But actually what I am expecting is to return all Person nodes. How to tell the person index of Elasticsearch to point on Person nodes in Neo4j?

These are the settings of person index:

curl XGET http://localhost:7474/person/_settings?pretty

{
  "person" : {
    "settings" : {
      "index" : {
        "gas" : {
          "enable" : "false",
          "neo4j" : {
            "user" : "neo4j",
            "hostname" : "http://localhost:7474",
            "password" : "neo4j."
          }
        },
        "creation_date" : "1495006378748",
        "number_of_shards" : "5",
        "number_of_replicas" : "1",
        "uuid" : "PuNk1GskRrW5_1BbGGWPiA",
        "version" : {
          "created" : "2030299"
        }
      }
    }
  }
}

Solution

Adding the GraphAidedSearch settings to the ES index requires this index to be created, as the error mentioned.

You can do this by issuing the following CURL request :

CURL -XPUT "http://localhost:9200/Person

Using an Elasticsearch extension (plugin) that allow you to enhance search results with graph-based recommendations requires that you have at least a minimum understanding of Elasticsearch itself.

I would highly recommend that you start by learning the basics of Elasticsearch before using any plugin.