Open
Description
Elastic.Clients.Elasticsearch version: 8.11.0
Elasticsearch version: 8.8.2
Operating system version: Windows 10
Description of the problem including expected versus actual behavior:
Elasticsearch-net can't deserialize the custom_token_chars field in an edge_ngram tokenizer because the field is defined as a string? when it is in fact an array of strings.
{"The JSON value could not be converted to System.String. Path: $.custom_token_chars | LineNumber: 0 | BytePositionInLine: 65."}
"settings": {
"index": {
"max_ngram_diff": "30",
"routing": {
"allocation": {
"include": {
"_tier_preference": "data_content"
}
}
},
"refresh_interval": "45s",
"number_of_shards": "1",
"provided_name": "index",
"creation_date": "1700171133856",
"analysis": {
"analyzer": {
"autocomplete": {
"filter": [
"trim",
"lowercase"
],
"tokenizer": "autocomplete"
},
"autocomplete_search": {
"filter": [
"trim",
"lowercase"
],
"tokenizer": "keyword"
},
"common_analyzer": {
"filter": [
"trim",
"lowercase"
],
"type": "custom",
"tokenizer": "keyword"
}
},
"tokenizer": {
"autocomplete": {
"token_chars": [
"letter",
"digit",
"custom"
],
"custom_token_chars": [
"()[]-&+"
],
"min_gram": "1",
"type": "edge_ngram",
"max_gram": "20"
}
}
},
"number_of_replicas": "1",
"uuid": "uuid",
"version": {
"created": "8080299"
}
}
}`
Steps to reproduce:
- Attempt to get the IndexState for an index with an edge_ngram tokenizer.
Expected behavior
It deserializes properly.