Enrich source index compression

Elastic Stack Elasticsearch
1

I have a very large source index for an enrich policy that stores two sparse_vector fields, which is almost 500gb on disk (no replicas).

Here's the source index mappings:

{
  "mappings": {
    "dynamic_templates": [],
    "properties": {
      "embedding_one": {
        "type": "sparse_vector"
      },
      "embedding_two": {
        "type": "sparse_vector"
      }
      "doc_id": {
        "type": "keyword"
      },
      "is_truncated": {
        "type": "boolean"
      }
    }
  }
}

And here's the enrich policy:

{
  "match": {
    "indices": "source_index_1",
    "match_field": "doc_id",
    "enrich_fields": ["embedding_one", "embedding_two"]
  }
}

What can I do from Tune for Disk Usage to the source index? In other words, what features does an enrich policy not require of its source index?

For example, does the enrich policy require the source index to keep _source enabled? Can I set "index": false on the sparse vectors? What about on the "doc_id"?