Stemmer class not found for light_bengali

I am getting an error while creating an index contains light_bengali stemmer which is supported by es. Stemmer token filteredit

es version: 7.2.0, I tried 7.6.1 too but got the same error

mapping

        {
          "settings": {
            "number_of_shards": 1,
            "number_of_replicas": 0,
            "refresh_interval": "10s",
            "analysis": {
              "filter": {
                "my_stop": {
                  "type": "stop",
                  "stopwords": "_bengali_"
                },
                "my_stemmer": {
                  "type": "stemmer",
                  "language": "light_bengali"
                }
              },
              "analyzer": {
                "bengali_analyzer": {
                  "tokenizer": "standard",
                  "filter": [
                    "lowercase",
                    "indic_normalization",
                    "bengali_normalization",
                    "my_stop",
                    "my_stemmer"
                  ]
                }
              }
            }},
    ```
and here is the error
    ```
        "type" : "illegal_argument_exception",
            "reason" : "Invalid stemmer class specified: Light_bengali",
            "caused_by" : {
              "type" : "class_not_found_exception",
              "reason" : "org.tartarus.snowball.ext.Light_bengaliStemmer"
            }
    ```

This looks like a documentation issue, where this has been copy/pasted from somewhere else. Would you mind opening a github issue?