[Python] Validating ES query_string upfront, namely without connecting to an Elasticsearch...

I am looking for a Python library (if any) that could help validate the query_string field of Elasticsearch queries1 upfront, namely without connecting to an Elasticsearch server and without having to define any custom validation logic. 1I mean queries of the kind

{
  "query": {
    "query_string": {
      "query": query_string
    }
  }
}


To give a clearer picture, I'd like to implement a simple utility that - based on a library of the kind described above - can help me build and let pass tests like the ones below:

@pytest.mark.parametrize(
  "data, expectation",
  [
    (
      ['("Extended Reality" AND "Mixed Reality")'],
      None
    ),
    # missing closed parenthesis
    (
      ['("Extended Reality" AND "Mixed Reality"'],
      pytest.raises(ValueError)
    ),
    # empty query
    ([''], pytest.raises(ValueError)),
    # invalid query
    (
      ['("Extended Reality" AND'],
      pytest.raises(ValueError)
    ),
  ],
)


I've tried to follow the suggestions of LLM-based tools which, however, have been proven wrong. For instance, I've tried out different suggested options exploiting elasticsearch_dsl and either the class Query

Query(query_str).to_dict() != {None: {}}


or classes Q and Search in combination

query = Q('query_string', query=query_string)
search=Search().query(query)
search.to_dict() != {None: {}}


, but none of them do work (as indeed elasticsearch_dsl is not meant for validation). Thank you!

Continue reading...