reindex_with_elasticsearch_helper.py

def transform_some_stuff(data, idx_and_doc_type):

    # This is mostly copied out of expand_action in the elasticserarch helpers thing -
    # which is why there's a bunch of keys I don't use in there.
    # But I'm just leaving them in anyways
    data = data.copy()
    op_type = data.pop('_op_type', 'index')
    action = {op_type: {}}
    for key in ('_index', '_parent', '_percolate', '_routing', '_timestamp',
                '_ttl', '_type', '_version', '_version_type', '_id', '_retry_on_conflict'):
        if key in data:
            action[op_type][key] = data.pop(key)

    action[op_type]['_type'] = idx_and_doc_type

    source = data['_source']

    # do stuff to the source document here

    return action, source



transform_some_stuff_with_version = partial(transform_some_stuff, idx_and_doc_type=index_config['idx_and_doc_type'])
res = elasticsearch.helpers.reindex(es, source_index_name, destination_index_name, chunk_size=500,
        bulk_kwargs={'expand_action_callback': transform_some_stuff_with_version})