elasticsearch-py driver is not filtering data properly when aggregating

Question

I faced a weird with elasticsearch python drivers and would like if someone can explain it to me! The below code works directly from cURL but doesn't work with python-requests or elasticsearch-py, strangely, it works when I switch to pyelasticsearch library! The details are:

I have a type called MY_TYPE that has a nested object MY_NESTED_FIELD and a child document MY_CHILD_TYPE. I'm trying to do term facet aggregation on the nested attributes based on filters applied to the MY_TYPE and MY_CHILD_TYPE types. The query looks like

{
  "query": {
    "filtered": {
      "filter": {
        "has_child": {
          "query": {
            "range": {
              "CHILD_FIELD": {
                "gte": 0.5
              }
            }
          },
          "type": "MY_CHILD_TYPE"
        }
      }
    }
  },    

  "aggs": {
    "aggregation_results": {
      "aggs": {
        "boards": {
          "terms": {
            "field": "MY_NESTED_FIELD.KEY",
            "size": 100
          },
          "aggs": {
            "MY_RANGES": {
              "range": {
                "ranges": [
                  {
                    "to": 0.5,
                    "from": 0
                  },
                  {
                    "to": 0.8
                    "from": 0.5
                  }
                ],
                "field": "MY_NESTED_FIELD_PATH.VALUE"
              }
            }
          }
        }
      },
      "nested": {
        "path": "MY_NESTED_FIELD_PATH"
      }
    }
  }
}

When I run this query against elasticsearch directly (using cURL or head plugin) it filters the parent and returns aggregations based on correct results. However, when I try it from the python script, it runs successfully but returns wrong data (it returns facets from all the documents without applying the filter)

I have tried:

• cURL: Works!
• ElasticSearch's HEAD plugin: Works!
• python-requests version 2.8.1: Did not work!
• elasticsearch-py api versions 1.4.0 and 2.1.0: Did not work!
• pyelasticsearch version 1.4: Works!

The code snippets for elasticsearch-py is:

from elasticsearch import Elasticsearch
es   = Elasticsearch('HOST:PORT')
data = es.search(index='INDEX_NAME', doc_type='MY_TYPE', body=payload, q='*:*', size=0)

When using python-requests, the code was:

import requests
url = 'http://ES_HOST:ES_PORT/ES_INDEX/ES_TYPE/_search'
params = {'size':0, 'q':'*:*'}
data   = requests.post(url, params=params, data=json.dumps(payload)).json()

My elastic search version is:

{
  "version": {
    "number": "1.4.4",
    "build_hash": "c88f77ffc81301dfa9dfd81ca2232f09588bd512",
    "build_timestamp": "2015-02-19T13:05:36Z",
    "build_snapshot": false,
    "lucene_version": "4.10.3"
  }
}

So my questions are:

1. Is this the best way to write this query?
2. Is there an explanation for why elasticsearch-py is acting strangely?
3. Is there a fix for this on elasticsearch-py?