SSE Client for Wikimedia

wiki-sse.py

#!/usr/bin/env python3

import json, time, sys, os
from sseclient import SSEClient as EventSource

from google.cloud import pubsub_v1

project_id = os.environ['PROJECT']
topic_name = os.environ['TOPIC']

publisher = pubsub_v1.PublisherClient()
topic_path = publisher.topic_path(project_id, topic_name)

futures = dict()

url = 'https://stream.wikimedia.org/v2/stream/recentchange'

keys = ['id', 'timestamp', 'user', 'title']

for event in EventSource(url):
    if event.event == 'message':
        try:
            change = json.loads(event.data)
            changePub = {k: change.get(k, 0) for k in keys}
        except ValueError:
            pass
        else:
            payloadJson = json.dumps(changePub).encode('utf-8')
            future = publisher.publish(
                   topic_path, data=payloadJson)
            futures[payloadJson] = future

while futures:
    time.sleep(5)