Elasticsearch introduction

gistfile1.txt

# Kibana version should match Elasticsearch version
GET /
 
DELETE services_v1
DELETE services_v2
DELETE resources
DELETE services_catalog_v1
 
 
 
 
 
 
 
 
 
# Get the list of all indices
GET _cat/indices
GET _cat/indices?v

# Create a new index. No mapping is required
PUT services_catalog_v1

# Create an alias "services"︙
POST _aliases/
{
  "actions": [
    {
      "add": {
        "alias": "services",
        "index": "services_catalog_v1"
      }
    }
  ]
}
# It is the same as 
PUT services_catalog_v1/_alias/services


GET _cat/indices?v
GET _cat/aliases

# Mapping and index name work the same
GET services
# is the same as 
GET services_catalog_v1







# Add some documents.
# No mapping, no ID needed.
# Every POST creates a new document if ID is not provided in the URL
POST services/service
{
  "name": "Fast service",
  "description": "<em>The service is providing fast access to <b>anything good</b> you want<em>",
  "created": "2017-11-02",
  "response_time": {
    "gte": 0.001,
    "lte": 0.01
  }
}









# The mapping is automatically created
GET services/
GET services/_mapping




# List documents, see the ID was autogenerated
GET services/_search


# If ID is provided in the URL, then the document created once and updated after
POST services/service/20
{
  "name": "Slow service",
  "description": "<em>This service is good in returing a data slowly<em>"
}





# Get the document including an index name, document type, version of the document
GET services/service/20

# Returns only the document itself
GET services/service/20/_source



# You can access the index by the index name and by the alias
GET services_catalog_v1/service/_search
GET services/service/_search

GET services/_search
{
  "query": {"match_all": {}}
}






# Languauge specific analysers remove stop words, trim words,  lowercase
GET _analyze
{"analyzer": "english", 
"text": "This service is GOOD in returing a data slowly"}




GET _analyze
{"analyzer": "english",
"text": "is not good"}




GET _analyze
{"analyzer": "german",
"text": "der Winter ist diesmal nicht eingetreten"}


# Mappings
# Available types: text, date, numeric (integer, float, double, etc.), date, boolean, range (integer_range, float_range, date_range, etc.), geo-point, geo-shapes, JSON object, nested JSON object.
# Arrays are not a separate type, any field can have multiple values. If you write script-queries in Painless you have to check the type.

# Create an index with a mapping
PUT services_v2
{
  "mappings": {
    "service": {
      "properties": {
        "description": {
          "type": "text",
          "analyzer": "english"
        },
        "created": {
          "type": "date"
        },
        "response_time":{
          "type": "float_range"
        }
        }
      }
    }
  }
  
  
  
  
# Swap indices
POST /_aliases
{
    "actions": [
        { "remove": { "index": "services_catalog_v1", "alias": "services" }},
        { "add":    { "index": "services_v2", "alias": "services" }}
    ]
}

GET services_catalog_v1/service/_mapping
GET services_v2/service/_mapping
GET services/_search



# Reindexing 
POST /_reindex
{"source": {"index": "services_catalog_v1"}, "dest": {"index":"services"}}
POST services/_search

# Remove the index
DELETE services_catalog_v1


# Creating documents
POST services/service/22
{
  "name": ["Fast service"],
  "description": "This service is absolutely the best service comparing to all other services"
}
GET services/service/22

# documents are immutable, they cannot be changed, only replaced
POST services/service/20
{
  "name": "Slow service",
  "description": "<em>This service is good in returing a data slowly<em>"
}



GET services/service/20
DELETE services/service/20
POST services/service/20/_update
{
  "doc": {
    "created": "1998-04-13",
    "response_time": {
      "gte": 0.5,
      "lte": 100
    }
  }
}

# Get all service documents
GET services/service/_search


# Try to search with "good", "is good", "is not good"
GET services/service/_search
{
  "query": {
    "match": {
      "description": "services is good"
    }
  }
}

# Try to search with "match_phrase"
GET services/service/_search
{
  "query": {
    "match_phrase": {
      "description": "is good"
    }
  }
}
GET services/service/_search
{
  "query": {
    "match_phrase": {
      "description": "service is provided"
    }
  }
}
GET services/service/_search
{
  "query": {
    "match_phrase": {
      "description": "good service"
    }
  }
}
GET services/service/_search
{
  "query": {
    "match": {
      "description": "good service"
    }
  }
}

# Fuzzy search
GET services/service/_search
{
  "query": {
    "fuzzy": {
      "description": "campari"
    }
  }
}
GET services/service/_search
{
  "query": {
    "fuzzy": {
      "description": "rest"
    }
  }
}

# Search by range
GET services/service/_search
{
  "query": {
    "range": {
      "created": {"gte": "1918-01-01", "lte": "now"}
    }
  }
}


# Search by range
GET services/service/_search
{
  "query": {
    "range": {
      "created": {"gte": "2018-01-01", "lte": "now"}
    }
  }
}

# Combine queries + highlight description
GET services/service/_search
{
  "query": {
    "bool": {
      "should": [
        {"fuzzy": {"description": "serpice"}},
        {"range": {"created": {"gte": "2000-01-01", "lte": "now"}}}
        ]
    }
  },
      "highlight" : {
        "fields" : {
            "description" : {}
        }
    }
}

# Customize highlighting
GET services/service/_search
{
  "query": {
    "match": {
      "description": "service is slow"
    }
  },
      "highlight" : {
        "pre_tags" : ["<b>"],
        "post_tags" : ["</b>"],
        "fields" : {
            "description" : {"fragment_size" : 35}
        }
    }
}
  
  
# Analyzer is a combination of filters and tokenizers
# https://www.elastic.co/guide/en/elasticsearch/reference/5.4/analysis-lang-analyzer.html#english-analyzer
# Standard english analyser trims words and remove stop words
GET _analyze
{"analyzer": "english",
"text": "<h1>There are no good drivers</h1>"}

# stop filter removes stop words: is, not, are ...
# snowball filter stems words: drivers -> driver
GET _analyze
{
  "char_filter": [
    "html_strip"
  ],
  "tokenizer": "standard", 
  "filter": ["stop", "snowball", "lowercase"], 
  "text": "<h1>There are no GOOD drivers impossible revolution disable</h1>"
}

# Scoring
# Search index takes frequency and matches into account
POST services/service/_search
{
  "query": {
    "match": {
      "description": "service"
    }
  }
}


# Explain search
POST services/service/_search?explain=true
{
  "query": {
    "match": {
      "description": "service"
    }
  }
}



# NGram, Trigram, Edge Gram
POST _analyze
{
  "tokenizer": {
    "type": "ngram",
    "min_gram": "3",
    "max_gram": "6",
    "token_chars": [
      "letter"
    ]
  },
  "filter": [
    "lowercase"
  ],
  "char_filter": ["html_strip"], 
  "text": [
    "1. <b>McDonalds</b> is the most popular restaurant in the world."
  ]
}


POST _analyze
{
  "tokenizer": {
    "type": "edge_ngram",
    "min_gram": "3",
    "max_gram": "4",
    "token_chars": [
      "letter"
    ]
  },
  "filter": [
    "lowercase"
  ],
  "char_filter": ["html_strip"], 
  "text": [
    "1. <b>McDonalds</b> is the most popular restaurant in the world."
  ]
}