[Feature Request] Implements fingerprint ingest processor

[gaobinlong]

Is your feature request related to a problem? Please describe

Currently we have community_id ingest processor which is used to generate community ID flow hash for network flow tuples based on the community id hash algorithm, but for common data such as application log or e-commerce data, we can also introduce a new type of ingest processor which can generate hash value based on part of the fields or all fields in a document, just like content hash, the fingerprint for each document can be used to deduplicate the documents and collapse search results.

The usage of the new fingerprint ingest processor could be:

"processors": [
      {
        "fingerprint": {
          "fields": ["foo", "bar"],
          "target_field": "fingerprint"
        }
      }
    ]

or

"processors": [
      {
        "fingerprint": {
          "include_all": true,
          "target_field": "fingerprint"
        }
      }
    ]

, after executing the processor, a new field fingerprint will be added to each document, then users can use the value of that field to deduplicate documents:

1. check if there are duplicated documents based on the fingerprint of each document
GET test1/_search
{
  "size": 0,
  "aggs": {
    "test": {
      "terms": {
        "field": "fingerprint",
        "min_doc_count": 2
      }
    }
  }
}
, the result is:
...
"aggregations": {
    "test": {
      "doc_count_error_upper_bound": 0,
      "sum_other_doc_count": 0,
      "buckets": [
        {
          "key": "MspgpPqOACPsB5VvjDbn1PdaClo=",
          "doc_count": 2
        }
      ]
    }
  }
, then user knows that there are 2 documents having same fingerprint, they may decide to delete one of them if it's not as expected.

, another use case is for collapsing search results:

GET test1/_search
{
  "collapse": {
    "field": "fingerprint"                
  }
}
, the search hits will only have one document if there're more than one hit documents containing the same fingerprint.

Describe the solution you'd like

Add a new ingest processor which can generate fingerprint for the incoming document.

Related component

Indexing

Describe alternatives you've considered

Generate the fingerprint in client side, which is not friendly for users.

Additional context

No response