We input the key word to get the relative tweets with tweepy interface and show them on the web page, then we adopt the algorithm proposed by Aggarwal et.al. to detect the event and get the clusters. For the five most recent clusters, we choose two most relative tweets with LexRank alogrithm to show on the web page. We adopt the flask framework and redis to implement the server-send-event technology of the website with gunicorn HTTP server.
test1/get_tweets.py : get tweets by tweepy interface
test1/get_event_t : get event
test1/t2.py : the main method
test1/template/ : web template
test1/static/ : web rendering
embedding/google_vectors/s1.py : get embedding with rpc server