lucene/solr/example/films/film_data_generator.py

117 lines
3.7 KiB
Python

# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements. See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
This will generate a movie data set of 1100 records.
These are the first 1100 movies which appear when querying the Freebase of type '/film/film'.
Here is the link to the freebase page - https://www.freebase.com/film/film?schema=
Usage - python3 film_data_generator.py
"""
import csv
import copy
import json
import codecs
import datetime
import urllib.parse
import urllib.request
import xml.etree.cElementTree as ET
from xml.dom import minidom
MAX_ITERATIONS=10 #10 limits it to 1100 docs
# You need an API Key by Google to run this
API_KEY = '<insert your Google developer API key>'
service_url = 'https://www.googleapis.com/freebase/v1/mqlread'
query = [{
"id": None,
"name": None,
"initial_release_date": None,
"directed_by": [],
"genre": [],
"type": "/film/film",
"initial_release_date>" : "2000"
}]
def gen_csv(filmlist):
filmlistDup = copy.deepcopy(filmlist)
#Convert multi-valued to % delimited string
for film in filmlistDup:
for key in film:
if isinstance(film[key], list):
film[key] = '|'.join(film[key])
keys = ['name', 'directed_by', 'genre', 'type', 'id', 'initial_release_date']
with open('films.csv', 'w', newline='', encoding='utf8') as csvfile:
dict_writer = csv.DictWriter(csvfile, keys)
dict_writer.writeheader()
dict_writer.writerows(filmlistDup)
def gen_json(filmlist):
filmlistDup = copy.deepcopy(filmlist)
with open('films.json', 'w') as jsonfile:
jsonfile.write(json.dumps(filmlist, indent=2))
def gen_xml(filmlist):
root = ET.Element("add")
for film in filmlist:
doc = ET.SubElement(root, "doc")
for key in film:
if isinstance(film[key], list):
for value in film[key]:
field = ET.SubElement(doc, "field")
field.set("name", key)
field.text=value
else:
field = ET.SubElement(doc, "field")
field.set("name", key)
field.text=film[key]
tree = ET.ElementTree(root)
with open('films.xml', 'w') as f:
f.write( minidom.parseString(ET.tostring(tree.getroot(),'utf-8')).toprettyxml(indent=" ") )
def do_query(filmlist, cursor=""):
params = {
'query': json.dumps(query),
'key': API_KEY,
'cursor': cursor
}
url = service_url + '?' + urllib.parse.urlencode(params)
data = urllib.request.urlopen(url).read().decode('utf-8')
response = json.loads(data)
for item in response['result']:
del item['type'] # It's always /film/film. No point of adding this.
try:
datetime.datetime.strptime(item['initial_release_date'], "%Y-%m-%d")
except ValueError:
#Date time not formatted properly. Keeping it simple by removing the date field from that doc
del item['initial_release_date']
filmlist.append(item)
return response.get("cursor")
if __name__ == "__main__":
filmlist = []
cursor = do_query(filmlist)
i=0
while(cursor):
cursor = do_query(filmlist, cursor)
i = i+1
if i==MAX_ITERATIONS:
break
gen_json(filmlist)
gen_csv(filmlist)
gen_xml(filmlist)