How to implement good search on product name in Elasticsearch. http://igor.kupczynski.info/2014/10/22/search-for-product-name.html

analysis.yaml

analyzer:
    generic_name_analyzer:
        type: "custom"
        tokenizer: "icu_tokenizer"
        filter: ["word_split", "icu_folding", "english_stop"]
    trigram_name_analyzer:
        type: "custom"
        tokenizer: "icu_tokenizer"
        filter: ["icu_folding", "english_stop", "trigram_filter"]

filter:
    word_split:
        type: "word_delimiter"
        preserve_original: 1
    english_stop:
        type: "stop"
        stopwords: "_english_"
    trigram_filter:
        type: "ngram"
        min_gram: 3
        max_gram: 3

documents.yaml

_type: file
fileName: TheSmallYellowDog.txt
---
_type: file
fileName: black cat.txt
---
_type: file
fileName: สวัสดี ผมมาจากกรุงเทพฯ.png

mappings.yaml

file: 
    properties: 
        fileName: 
            type: multi_field
            fields: 
                fileName: 
                    type: string
                    analyzer: generic_name_analyzer
                trigram: 
                    type: string
                    analyzer: trigram_name_analyzer

play.sh

#!/bin/bash

export ELASTICSEARCH_ENDPOINT="http://localhost:9200"

# Create indexes

curl -XPUT "$ELASTICSEARCH_ENDPOINT/play" -d '{
    "settings": {
        "analysis": {
            "analyzer": {
                "generic_name_analyzer": {
                    "type": "custom",
                    "tokenizer": "icu_tokenizer",
                    "filter": [
                        "word_split",
                        "icu_folding",
                        "english_stop"
                    ]
                },
                "trigram_name_analyzer": {
                    "type": "custom",
                    "tokenizer": "icu_tokenizer",
                    "filter": [
                        "icu_folding",
                        "english_stop",
                        "trigram_filter"
                    ]
                }
            },
            "filter": {
                "word_split": {
                    "type": "word_delimiter",
                    "preserve_original": 1
                },
                "english_stop": {
                    "type": "stop",
                    "stopwords": "_english_"
                },
                "trigram_filter": {
                    "type": "ngram",
                    "min_gram": 3,
                    "max_gram": 3
                }
            }
        }
    },
    "mappings": {
        "file": {
            "properties": {
                "fileName": {
                    "type": "multi_field",
                    "fields": {
                        "fileName": {
                            "type": "string",
                            "analyzer": "generic_name_analyzer"
                        },
                        "trigram": {
                            "type": "string",
                            "analyzer": "trigram_name_analyzer"
                        }
                    }
                }
            }
        }
    }
}'

# Index documents
curl -XPOST "$ELASTICSEARCH_ENDPOINT/_bulk?refresh=true" -d '
{"index":{"_index":"play","_type":"file"}}
{"fileName":"TheSmallYellowDog.txt"}
{"index":{"_index":"play","_type":"file"}}
{"fileName":"black cat.txt"}
{"index":{"_index":"play","_type":"file"}}
{"fileName":"สวัสดี ผมมาจากกรุงเทพฯ.png"}
'

# Do searches

curl -XPOST "$ELASTICSEARCH_ENDPOINT/_search?pretty" -d '
{
    "query": {
        "match": {
            "fileName": "ผม"
        }
    }
}
'

curl -XPOST "$ELASTICSEARCH_ENDPOINT/_search?pretty" -d '
{
    "query": {
        "bool": {
            "should": [
                {
                    "match": {
                        "fileName": {
                            "query": "yellowish",
                            "boost": 3
                        }
                    }
                },
                {
                    "match": {
                        "fileName.trigram": {
                            "query": "yellowish",
                            "minimum_should_match": "50%",
                            "boost": 1
                        }
                    }
                }
            ]
        }
    }
}
'

play.yaml

# Auto generated by Found's Play-tool at 2015-01-20T02:17:23+08:00
version: 0
title: Searching for Product Name in Elasticsearch
description: "How to implement good search on product name in Elasticsearch. http://igor.kupczynski.info/2014/10/22/search-for-product-name.html"

searches.yaml

---
query: 
    match:
        fileName: "ผม"     
---
query:
    bool: 
        should: 
            - match: 
                  fileName:
                    query: "yellowish"
                    boost: 3
            - match: 
                  fileName.trigram: 
                      query: "yellowish"
                      minimum_should_match: "50%"
                      boost: 1