Create custom analyzer filter for a index in py-elasticsearch-dsl

i'm working with py-elasticsearch-dsl for my master, i'm creating a index of title documents in a corpus of turkish titles, and i need implements a custom lowercase analyzer for turkish language: https://www.elastic.co/guide/en/elasticsearch/reference/current/analysis-lowercase-tokenfilter.html#analysis-lowercase-tokenfilter

I'm trying do it with this:

turkish = analysis.token_filter('turkish_lowercase', type="lowercase", language="turkish")


turkish_lowercase = analyzer('turkish_lowercase',
    type = "custom",
    tokenizer="standard",
    filter=["turkish_lowercase"],
)

class Document(DocType):
    # title = Text()
    query = Percolator(
        analyzer=turkish_lowercase,
        filter=turkish
    )    # query is a percolator

    class Meta:
        index = 'titles' # index name
        doc_type = '_doc'

    def save(self, **kwargs):
        return super(Document, self).save(**kwargs)

But i'm getting this error:

python percolator.py                                        1 ↵  1736  17:37:54 
PUT http://localhost:9200/title-index [status:400 request:0.004s]
Traceback (most recent call last):
  File "percolator.py", line 55, in <module>
    Document.init()
  File "/home/salahaddin/Proyectos/Works/seminer/lib/python3.6/site-packages/elasticsearch_dsl/document.py", line 161, in init
    cls._doc_type.init(index, using)
  File "/home/salahaddin/Proyectos/Works/seminer/lib/python3.6/site-packages/elasticsearch_dsl/document.py", line 85, in init
    self.mapping.save(index or self.index, using=using or self.using)
  File "/home/salahaddin/Proyectos/Works/seminer/lib/python3.6/site-packages/elasticsearch_dsl/mapping.py", line 116, in save
    return index.save()
  File "/home/salahaddin/Proyectos/Works/seminer/lib/python3.6/site-packages/elasticsearch_dsl/index.py", line 219, in save
    return self.create()
  File "/home/salahaddin/Proyectos/Works/seminer/lib/python3.6/site-packages/elasticsearch_dsl/index.py", line 203, in create
    self.connection.indices.create(index=self._name, body=self.to_dict(), **kwargs)
  File "/home/salahaddin/Proyectos/Works/seminer/lib/python3.6/site-packages/elasticsearch/client/utils.py", line 76, in _wrapped
    return func(*args, params=params, **kwargs)
  File "/home/salahaddin/Proyectos/Works/seminer/lib/python3.6/site-packages/elasticsearch/client/indices.py", line 91, in create
    params=params, body=body)
  File "/home/salahaddin/Proyectos/Works/seminer/lib/python3.6/site-packages/elasticsearch/transport.py", line 314, in perform_request
    status, headers_response, data = connection.perform_request(method, url, params, body, headers=headers, ignore=ignore, timeout=timeout)
  File "/home/salahaddin/Proyectos/Works/seminer/lib/python3.6/site-packages/elasticsearch/connection/http_urllib3.py", line 163, in perform_request
    self._raise_error(response.status, raw_data)
  File "/home/salahaddin/Proyectos/Works/seminer/lib/python3.6/site-packages/elasticsearch/connection/base.py", line 125, in _raise_error
    raise HTTP_EXCEPTIONS.get(status_code, TransportError)(status_code, error_message, additional_info)
elasticsearch.exceptions.RequestError: TransportError(400, 'illegal_argument_exception', 'Custom Analyzer [turkish_lowercase] failed to find filter under name [turkish_lowercase]')

So, what is the correct way to do this?

Thank you

For create a custom filter we can use token_filter:

turkish = analysis.token_filter('turkish_lowercase', type="lowercase", language="turkish")

We are creating a new lower_case filter for turkish language. Now, we need to create the analyzer:

turkish_lowercase = analyzer('turkish_lowercase',
    type = "custom",
    tokenizer="standard",
    filter=[turkish],
)

We put turkish filter directly in the filter parameter, nothing more; we can review the result dictionary with the get_definition function either filter as analyzer.

Finally we apply that analyzer in our Document:

class Document(DocType):
    title = Text(
        analyzer=turkish_lowercase,
        # filter=turkish
        )
    query = Percolator(
    )    # query is a percolator

    class Meta:
        index = 'titles' # index name
        doc_type = '_doc'

    def save(self, **kwargs):
        return super(Document, self).save(**kwargs)

We will get the next result:

{
   "titles":{
      "aliases":{

      },
      "mappings":{
         "_doc":{
            "properties":{
               "query":{
                  "type":"percolator"
               },
               "title":{
                  "type":"text",
                  "analyzer":"turkish_lowercase"
               }
            }
         }
      },
      "settings":{
         "index":{
            "number_of_shards":"5",
            "provided_name":"titles",
            "analysis":{
               "filter":{
                  "turkish_lowercase":{
                     "type":"lowercase",
                     "language":"turkish"
                  }
               },
               "analyzer":{
                  "turkish_lowercase":{
                     "filter":[
                        "turkish_lowercase"
                     ],
                     "type":"custom",
                     "tokenizer":"standard"
                  }
               }
            },
            "number_of_replicas":"1",
         }
      }
   }
}