Elasticsearch in Python, is there a phrase suggestor?

Question

Following this link here. There is a concept called "Phrase Suggestor" which uses some N-Gram methods to give you a suggestions sort of like autocompletion. I was trying to see how to use the api that Python offers docs found here. But I could not find anything mentioning n-gram or phrase suggestor.

Does this method exist in the Python Elasticsearch API? I am aware of NLTK and the n-gram methods there.

Here is what I have.

First connecting, this block of code works fine

from elasticsearch import Elasticsearch
CLOUD_ID = 'My_deployment:...'
ELASTIC_PASSWORD = 'password'

es = Elasticsearch(cloud_id=CLOUD_ID,
    basic_auth=("elastic", ELASTIC_PASSWORD))

This second block does not work

text = 'noble prize'
suggest_dictionary = {"simple_phrase" : {
                      'text' : text,
                      "phrase" : {
                          "field" : "title.trigram"
                      }
                    }
                  }

query_dictionary = {'suggest' : suggest_dictionary}

res = es.search(
    index='test',
    body=query_dictionary)
print(res)

The error message is this

<ipython-input-29-05c434577314>:12: DeprecationWarning: The 'body' parameter is deprecated and will be removed in a future version. Instead use individual parameters.
  res = es.search(

---------------------------------------------------------------------------
NotFoundError                             Traceback (most recent call last)
<ipython-input-29-05c434577314> in <module>
     10 query_dictionary = {'suggest' : suggest_dictionary}
     11 
---> 12 res = es.search(
     13     index='test',
     14     body=query_dictionary)

~/anaconda3/lib/python3.8/site-packages/elasticsearch/_sync/client/utils.py in wrapped(*args, **kwargs)
    402                         pass
    403 
--> 404             return api(*args, **kwargs)
    405 
    406         return wrapped  # type: ignore[return-value]

~/anaconda3/lib/python3.8/site-packages/elasticsearch/_sync/client/__init__.py in search(self, index, aggregations, aggs, allow_no_indices, allow_partial_search_results, analyze_wildcard, analyzer, batched_reduce_size, ccs_minimize_roundtrips, collapse, default_operator, df, docvalue_fields, error_trace, expand_wildcards, explain, fields, filter_path, from_, highlight, human, ignore_throttled, ignore_unavailable, indices_boost, lenient, max_concurrent_shard_requests, min_compatible_shard_node, min_score, pit, post_filter, pre_filter_shard_size, preference, pretty, profile, q, query, request_cache, rescore, rest_total_hits_as_int, routing, runtime_mappings, script_fields, scroll, search_after, search_type, seq_no_primary_term, size, slice, sort, source, source_excludes, source_includes, stats, stored_fields, suggest, suggest_field, suggest_mode, suggest_size, suggest_text, terminate_after, timeout, track_scores, track_total_hits, typed_keys, version)
   3697         if __body is not None:
   3698             __headers["content-type"] = "application/json"
-> 3699         return self.perform_request(  # type: ignore[return-value]
   3700             "POST", __path, params=__query, headers=__headers, body=__body
   3701         )

~/anaconda3/lib/python3.8/site-packages/elasticsearch/_sync/client/_base.py in perform_request(self, method, path, params, headers, body)
    319                     pass
    320 
--> 321             raise HTTP_EXCEPTIONS.get(meta.status, ApiError)(
    322                 message=message, meta=meta, body=resp_body
    323             )

NotFoundError: NotFoundError(404, 'index_not_found_exception', 'no such index [test]', test, index_or_alias)

The answer provided states to use the PUT test to setup the index. Where? no idea... how? no idea... I am not familiar with that syntax and Python does not seem to be able to recognize it either.

Update

I was able to get it to work finally, but I am confused by the output

{'took': 1, 'timed_out': False, '_shards': {'total': 1, 'successful': 1, 'skipped': 0, 'failed': 0}, 'hits': {'total': {'value': 0, 'relation': 'eq'}, 'max_score': None, 'hits': []}, 'suggest': {'simple_phrase': [{'text': 'Hi, I need help', 'offset': 0, 'length': 15, 'options': []}]}}

Where is the recommendation? aka the autocompletion to finish the sentence?

Answer 1

You can first create index mapping with so you not need to depends on external python Ngram and Elasticsearch store field internally after generating NGram.

Index Mapping

PUT test
{
  "settings": {
    "index": {
      "number_of_shards": 1,
      "analysis": {
        "analyzer": {
          "trigram": {
            "type": "custom",
            "tokenizer": "standard",
            "filter": ["lowercase","shingle"]
          }
        },
        "filter": {
          "shingle": {
            "type": "shingle",
            "min_shingle_size": 2,
            "max_shingle_size": 3
          }
        }
      }
    }
  },
  "mappings": {
    "properties": {
      "title": {
        "type": "text",
        "fields": {
          "trigram": {
            "type": "text",
            "analyzer": "trigram"
          }
        }
      }
    }
  }
}

You can use below python code for Phrase Suggestor and provided body same as mentioned in documentation.

from elasticsearch import Elasticsearch
es = Elasticsearch()

text = 'noble prize'
suggest_dictionary = {"simple_phrase" : {
                      'text' : text,
                      "phrase" : {
                          "field" : "title.trigram"
                      }
                    }
                  }

query_dictionary = {'suggest' : suggest_dictionary}

res = es.search(
    index='test',
    body=query_dictionary)
print(res)

Update 1:

The answer provided states to use the PUT test to setup the index. Where? no idea... how? no idea... I am not familiar with that syntax and Python does not seem to be able to recognize it either.

put test is for creating index in Elasticsearch. So if you have kibana install then you can goto dev console and execute it. otherwise you can use same with curl command as well. If you have exsitig index then you can give your index name as well insted of test.

This will show how to use curl command for index creation.

This will show how to use python for creating index.