Configure a tokenizer with pyes

Question

I'm trying to configure one of my fields to use an edge ngram tokenizer. I'm trying to translate the following gist that I found (https://gist.github.com/1037563):

{
    "mappings": {
        "contact": {
            "properties": {
                "twitter": {
                    "type": "object",
                    "properties": {
                        "profile": {
                            "fields": {
                                "profile": {
                                    "type": "string",
                                    "analyzer": "left"
                                },
                                "reverse_profile": {
                                    "type": "string",
                                    "analyzer": "right"
                                }
                            },
                            "type": "multi_field"
                        }
                    }
                }
            }
        }
    },
    "settings": {
        "analysis": {
            "analyzer": {
                "left": {
                    "filter": [
                        "standard",
                        "lowercase",
                        "stop"
                    ],
                    "type": "custom",
                    "tokenizer": "left_tokenizer"
                },
                "right": {
                    "filter": [
                        "standard",
                        "lowercase",
                        "stop"
                    ],
                    "type": "custom",
                    "tokenizer": "right_tokenizer"
                }
            },
            "tokenizer": {
                "left_tokenizer": {
                    "side": "front",
                    "max_gram": 20,
                    "type": "edgeNGram"
                },
                "right_tokenizer": {
                    "side": "back",
                    "max_gram": 20,
                    "type": "edgeNGram"
                }
            }
        }
    }
}

I can see pyes supports the 'put_mapping' API, but this seems to wrap everything inside 'mappings'. I need to be able to pass the analyzer under a 'settings' key and can't work out how to.

Can anyone help?

Answer 1

You should be able to pass this structure as the second parameter of create_index.