Uax_url_email tokenizer not recognising valid emails with no dots on the email domain

Elastic Stack Elasticsearch
1

Using version 7.9. And having the following settings:

                      "email_filter": {
                            "type": "pattern_capture",
                            "preserve_original": "true",
                            "patterns": [
                                "([^@]+)",
                                "(\\p{L}+)",
                                "(\\d+)",
                                "@(.+)"
                            ]
                        }
                    },
                    "analyzer": {
                        "email_analyzer": {
                            "filter": [
                                "email_filter",
                                "lowercase"
                            ],
                            "tokenizer": "uax_url_email"
                        }
                    }

When doing an _analyze query:

{
    "analyzer": "email_analyzer",
    "text" : "myemail@gmail"
}

I get the following tokens:

{
    "tokens": [
        {
            "token": "myemail",
            "start_offset": 0,
            "end_offset": 7,
            "type": "<ALPHANUM>",
            "position": 0
        },
        {
            "token": "gmail",
            "start_offset": 8,
            "end_offset": 13,
            "type": "<ALPHANUM>",
            "position": 1
        }
    ]
}

Observe that it is not being recognised as EMAIL, it is showing ALPHANUM and I don't get the original token "myemail@gmail"

But when doing an _analyze query on an email with a dot on the domain:

{
    "analyzer": "email_analyzer",
    "text" : "myemail@gmail.com"
}

I get the expected response, being recognised as an EMAIL and "myemail@gmail.com" as the original token as well.

But why an email without a dot on the domain, which is valid, is not being recognised as EMAIL by the uax_url_email tokenizer?

2

The email tokenization comes out of Lucene, which uses strict domain name labels.

3

Thank you @Kathleen_DeRusso