Elastic Search Example

get_hits.py

import json
import argparse
import codecs
from  elastic_search import ElasticIndex

def process(args):
	url = args.url
	index = args.index
	es_type = args.type
	query = args.query
	output = args.output
	with codecs.open(query,'r','utf-8') as q:
		query = json.loads(q.read())

	e_index = ElasticIndex(url,index)
	hits = e_index.get_data(query,es_type)
	clean_entities = get_entities(hits)

	with codecs.open(output,'w','utf-8') as o:
		for e in clean_entities:
			o.write(u'{0}\n'.format(e))

def get_entities(hits):
	entities = []
	for h in hits["hits"]:
		entities.append(h["fields"]["relationships.parent.name"])
		# entities.extend(h["fields"]["aka"])
		entities.append(h["fields"]["name"])

	return set(entities)

def main():
	parser = argparse.ArgumentParser(description=
			'extract the entities elastic search')
	parser.add_argument('url', help='endpoint url', type = str)
	parser.add_argument('index',help='elastic index',type=str)
	parser.add_argument('type',help='elastic type')
	parser.add_argument('-q','--query',help='query to run', type=str)
	parser.add_argument('-o','--output',help='output file', type=str)

	args = parser.parse_args()

	process(args)


if __name__ == '__main__':
	main()