Search a list of names and categorizing each letter type - elasticsearch

Question

I want to index a large list of names using ES.
I want to distinguish between consonants and vowels in each word, and be able to search based on the position of each letter and if it is a consonant or a vowel.
So say the name like:
JOHN
I want to enter this:
CVCC
and when I run the search, JOHN should be in the result set.
Is it possible somehow to index names in elastic search such that I could index and then search them using the tokens C and V for vowel?
So somehow Elasticsearch will have to index the character types for each position for each word, how can this be done?

Answer 1

You can do it with pattern_replace char filters in a custom analyzer. Also, in my solution I have used a sub-field for the custom analyzer, thinking maybe that you will want other kinds of searches on the name field, the consonants-vowels one being only one of them.
DELETE test
PUT test
{
"settings": {
"analysis": {
"analyzer": {
"my_analyzer": {
"tokenizer": "keyword",
"char_filter": [
"replace_filter_lowercase_CONS",
"replace_filter_uppercase_CONS",
"replace_filter_lowercase_VOW",
"replace_filter_uppercase_VOW"
]
}
},
"char_filter": {
"replace_filter_lowercase_CONS": {
"type": "pattern_replace",
"pattern": "[b-df-hj-np-tv-z]{1}",
"replacement": "c"
},
"replace_filter_uppercase_CONS": {
"type": "pattern_replace",
"pattern": "[B-DF-HJ-NP-TV-Z]{1}",
"replacement": "C"
},
"replace_filter_lowercase_VOW": {
"type": "pattern_replace",
"pattern": "[aeiou]{1}",
"replacement": "v"
},
"replace_filter_uppercase_VOW": {
"type": "pattern_replace",
"pattern": "[AEIOU]{1}",
"replacement": "V"
}
}
}
},
"mappings": {
"test": {
"properties": {
"name": {
"type": "text",
"fields": {
"cons_vow": {
"type": "text",
"analyzer": "my_analyzer"
}
}
}
}
}
}
}
POST /test/test/1
{"name":"JOHN"}
POST /test/test/2
{"name":"Andrew"}
POST /test/test/3
{"name":"JOhn DOE"}
GET /test/_search
{
"query": {
"term": {
"name.cons_vow": {
"value": "CVCC"
}
}
}
}