Notes for my Elasticsearch tutorial at Boosterconf: http://boosterconf.no/talks/133/

Prereqiusites

Install Elasticsearch (http://www.elasticsearch.org/guide/reference/setup/installation.html) (requires Java)

Install the Elasticsearch head plugin (http://mobz.github.com/elasticsearch-head/)

Indexing, mapping and queries

Delete the index:

curl -X DELETE http://localhost:9200/talks

Chech the index:

curl -I "http://localhost:9200/talks"

Create initial index:

curl -X POST http://localhost:9200/talks -d '{"mappings":{},"settings":{"number_of_shards":1,"number_of_replicas":0}}'

Load some test data:

curl -s http://boosterconf.no/talks/194.json | xargs -0 echo | curl -X POST "http://localhost:9200/talks/talk/194" -d @-
curl -s http://boosterconf.no/talks/170.json | xargs -0 echo | curl -X POST "http://localhost:9200/talks/talk/170" -d @-
curl -s http://boosterconf.no/talks/177.json | xargs -0 echo | curl -X POST "http://localhost:9200/talks/talk/177" -d @-
curl -s http://boosterconf.no/talks/179.json | xargs -0 echo | curl -X POST "http://localhost:9200/talks/talk/179" -d @-

Or you could run load_data.sh

Run a simple query:

curl -XGET 'http://localhost:9200/talks/talk/_search?size=10&pretty' -d '{"query": {"query_string":{"query": "kafka"}}}'

With highlights:

curl -XGET 'http://localhost:9200/talks/talk/_search?size=10&pretty' -d '{
  "query": {
    "query_string": {
      "query": "kafka"
    }
  }, 
  "highlight": {
    "fields": {
      "title": {},
      "description":{}
    }
  }
}'

Search for HTML-markup:

curl -XGET 'http://localhost:9200/talks/talk/_search?size=10&pretty' -d '{
  "query": {
    "query_string": {
      "query": "description: <p>"
    }
  }, 
  "highlight": {
    "fields": {
      "title": {},
      "description":{}
    }
  }
}'

Delete the index:

curl -X DELETE http://localhost:9200/talks

Recreate index with our own custom analyzer:

curl -X PUT http://localhost:9200/talks -d '{
  "mappings":{},
  "settings":{
    "index": { 
      "number_of_shards":1,
      "number_of_replicas":0,
      "analysis": {
        analyzer: {
          "descriptionAnalyzer": {
            "type": "custom",
            "tokenizer": "standard",
            "filter": "standard",
            "char_filter": "html_strip"
          }
        }  
      }
    }  
  }  
}'

Test the analyzer:

curl -XGET 'localhost:9200/talks/_analyze?analyzer=standard' -d '<p>this is a test</p>'
curl -XGET 'localhost:9200/talks/_analyze?analyzer=descriptionAnalyzer' -d '<p>this is a test</p>'

Update the mapping:

curl -XPUT 'http://localhost:9200/talks/talk/_mapping' -d '{
    "talk" : {
        "properties" : {
            "description" : {"type" : "string", "store" : "yes", "analyzer": "descriptionAnalyzer"}
        }
    }
}'

Reload the test data, try to search for the html markup again.

Facets:

curl -XGET 'http://localhost:9200/talks/talk/_search?size=10&pretty' -d '{
  "query": {
    "query_string": {
      "query": "acceptance_status: accepted"
    }
  },
  "facets" : {
    "language" : {
        "terms" : { "field" : "language" }
    }
  }
}'

Filters:

curl -XGET 'http://localhost:9200/talks/talk/_search?size=10&pretty' -d '{
  "query": {
    "query_string": {
      "query": "acceptance_status: accepted"
    }
  },
  "filter" : {
    "term" : { "language" : "english" }
  }  
}'

Nested documents

Run ./load_nested_data.sh

curl -XGET localhost:9200/demo_cvs/cv/_search?pretty -d '{
  "query": {
    "query_string": {
      "query": "talks.title: type"
    }
  }
}'

curl -XGET localhost:9200/demo_cvs/cv/_search?pretty -d '{
  "query": {
    "query_string": {
      "query": "talks.title: \"tutorial type\""
    }
  }
}'

Add mapping:

curl -X DELETE http://localhost:9200/demo_cvs  

curl -X PUT http://localhost:9200/demo_cvs -d '{
  "mappings":{
    "cv" : {
      "properties" : {
        "talks" : {"type" : "nested"}
      }
    }
  }
}'

Try a the same a nested query:

curl -XGET localhost:9200/demo_cvs/cv/_search?pretty -d '{
  "query": { 
    "nested" : {
      "path" : "talks",
      "query" : {
        "bool" : {
          "must" : [
              {
                "query_string" : {"query" : "talks.title:\"tutorial type\""}
              }
            ]
        }
      }
    }
  }
}'

curl -XGET localhost:9200/demo_cvs/cv/_search?pretty -d '{
  "query": { 
    "nested" : {
      "path" : "talks",
      "query" : {
        "bool" : {
          "must" : [
              {
                "query_string" : {"query" : "talks.title:\"type driven\""}
              }
            ]
        }
      }
    }
  }
}'

Optionally, add a position_offset_gap to separate the fields

curl -X DELETE http://localhost:9200/demo_cvs

curl -X PUT http://localhost:9200/demo_cvs -d '{
  "mappings":{
    "cv" : {
      "properties" : {
        "talks" : {
          "type" : "nested", 
          "properties": {
            "title":{"type":"string","position_offset_gap":256}
          }
        }
      }
    }
  }
}'

Parent / Child documents

Load data and mappings by running ./load_parent_child_data

curl -X GET localhost:9200/parent_cvs/child_talk/_search -d '{
  "query": {
    "has_parent": {
      "type": "parent_cv",
      "query" : {
        "query_string": {
          "query": "name: super"
        }
      }
    }
  }
}'


curl -X GET localhost:9200/parent_cvs/parent_cv/_search -d '{
  "query": {
    "has_child": {
      "type": "child_talk",
      "query" : {
        "query_string": {
          "query": "title: \"elasticsearch\""
        }
      }
    }
  }
}'

Credits to http://www.spacevatican.org/2012/6/3/fun-with-elasticsearch-s-children-and-nested-documents/ for inspiration to this part of the tutorial

Libraries

https://github.com/karmi/tire