dump_mappings.py

""" Dump a set of user-friendly text files outlining all our ElasticSearch type mappings """

import json

type_props = []

def add_prop_and_remove_non_en_i18n(prop):
    """ Do some extra logic to remove superfluous i18n fields that aren't 'en' """
    if 'i18n' in prop and 'i18n.en' not in prop:
        return
    type_props.append(prop)

def handle_props(props, prefix=''):
    for prop, mapping in props.items():
        prop = prefix + prop

        if 'properties' in mapping:  # nested type, recursively get its properties too..
            handle_props(mapping['properties'], prop + '.')
        else:
            add_prop_and_remove_non_en_i18n(prop)

def handle_type(mapping):
    if 'properties' in mapping:
        handle_props(mapping['properties'])

if __name__ == '__main__':
    # mappings.txt is a file with JSON in it from querying ES for its mappings
    # e.g. curl -XGET 'http://localhost:9200/_mapping?pretty=true' > mappings.txt
    with open('mappings.txt') as f:
        mappings = json.load(f)

    for name, mapping in mappings['atom']['mappings'].items():
        fh = open('type_mappings/' + name + '.txt', 'w')
        fh.write('%s\n-----\n' % name)
        handle_type(mapping)
        for prop in sorted(type_props):
            fh.write(prop + '\n')
        fh.close()
        type_props = []

If the index you're pulling from isn't named "atom" you'll need to edit line 32 to reflect that. Further, this script is assuming a folder called type_mappings is created under the ./ (current working directory) so it can dump the text files there.