Elasticsearch 'failed to find analyzer' error & analyzer not shown by Settings API

Question

I'm trying to get my Elasticsearch indices to use the Porter stemming algorithm, but my custom analyzer isn't defined when I test with the _analyze endpoint.

I've looked at the ES documentation and similar questions on SO, and I'm not sure what the problem is. I tried using a separate PUT request for the settings when creating the index, but that had no effect.

This is how I create the mappings:

@staticmethod
    def make_mapping():
        mapping = {
            'settings': {
                'analysis':
                {
                    'analyzer': {
                        'porter_english': {
                            'type': 'custom',
                            'tokenizer': 'standard',
                            'stopwords': '_english_',
                            'filter': ['lowercase', 'porter_stem']
                            }
                        }
                    }
                },
            'mappings': {
                'properties': {
                    'published': {
                        'type': 'boolean'
                    },
                    'title': {
                        'type': 'text',
                        'analyzer': 'english'
                    },
                    'date': {
                        'type': 'date'
                    },
                    'description': {
                        'type': 'text',
                        'analyzer': {
                            'porter_english': {
                                'type': 'custom',
                                'tokenizer': 'standard',
                                'stopwords': '_english_',
                                'filter': ['lowercase', 'porter_stem']
                            }
                        }
                    },
                    'keywords': {
                        'type': 'text',
                        'analyzer': {
                            'porter_english': {
                                'type': 'custom',
                                'tokenizer': 'standard',
                                'stopwords': '_english_',
                                'filter': ['lowercase', 'porter_stem']
                            }
                        }
                    },
                    'price': {
                        'type': 'float'
                    }
                }
            }
        }
        return mapping

This is the function that creates the indexes from the mapping.

def init_elasticsearch():
    es = elasticsearch.Elasticsearch(['http://localhost:9200'])
    # idx_client = elasticsearch.client.IndicesClient(es)
    for i in searchables.included:
        index_name = camelTo_snake(i.__name__)
        index_m = i.make_mapping()
        index_uri = "{}/{}".format(current_app.config['ELASTIC_URI'], index_name)
        create_index = requests.put(index_uri, json=index_m)
        init_settings = requests.put(index_uri, json=index_m['settings'])

And this is all I get if I query the settings:

>>> g = requests.get(e + '/gallery_item/_settings')
>>> g.text
'{
    "gallery_item":{
        "settings":{
            "index":{
                "creation_date":"1564789941204",
                "number_of_shards":"1",
                "number_of_replicas":"1",
                "uuid":"SgkEBN4nTxWUCeSGWMwbGw",
                "version":{"created":"7020099"},
                "provided_name":"gallery_item"
             }
         }
    }
}'

I just need those two fields to use the porter_stem token filter.

Kevin Quinzel · Answer 1 · 2019-08-03T01:45:26.233

0

Have you considered using the Python ES client instead of using requests?

It lets you easily manipulate everything related to your cluster: From creating indices, setting properties, do queries, etc.

For your case, you can Set settings to your indices or Set mappings as well. Just use the correct args and everything should be fine.

Hope this is helpful :D

edited Aug 03 '19 at 01:45

answered Aug 03 '19 at 00:52

Kevin Quinzel

1,430
1
13
23

Yeah, thanks, I have used it a little bit--I found using requests was a little easier & more readable, but I'll see if it works using the Python ES client. – soda_hobart Aug 03 '19 at 01:00

score 0 · Accepted Answer · answered Aug 03 '19 at 12:55

I believe this part of the mappings is wrong

change this

'description': {
                        'type': 'text',
                        'analyzer': {
                            'porter_english': {
                                'type': 'custom',
                                'tokenizer': 'standard',
                                'stopwords': '_english_',
                                'filter': ['lowercase', 'porter_stem']
                            }
                        }
                    },

to

 'description': {
                 'type': 'text',
                  'analyzer': 'porter_english'    

  },

because you already defined analyzer in your settings. you just have to use it in the mappings

Thank you, this did the trick. I swear I actually tried to it set up like this originally, but it didn't work as expected, but maybe it was because I didn't do separate PUT requests for both the mapping and the settings? — soda_hobart, Aug 03 '19 at 15:24

Elasticsearch 'failed to find analyzer' error & analyzer not shown by Settings API

2 Answers2