def get(self):
self.response.out.write("""
<html>
<body>
<h2>Welcome!</h2>
This is assignment3 for Columbia Univ 6998 Cloud Computing. By Kina Winoto (ksw2123) and Yongxu Zhang (yz2419).</br>
It is built using Google App Engine, Python SDK.
Note that the assignment specification was a little vague and where there was ambiguity, we made our own decisions.</br>
Thus, here's a little about it to clear up some confusion: </br>
Enter a topic below. </br>
Returned, is the "buzz" (the most significant tweet) about that topic. We calculate that by: </br>
<ol><li>Counting the frequency of each word in all of the tweets associated with
your topic, excluding words such as "a", "the", etc..</li>
<li>Using that frequency, we give each tweet a score. The more words with non-zero frequencies
that a tweet has, the higher the score.</li>
<li>The buzz-tweet is the tweet with the highest score.</li>
</ol>
We use DataStore to store all of the tweets and buzz-tweets associated with a topic.</br>
We use memcache to store each topic's buzz-tweet. (It is kept for only 1 minute within memcache so
that buzzes are always up to date) </br>
You can click on Tweet Contents to see all of the tweets about that topic (non-formatted). </br>
<h3>Please enter the topic:</h3>
<form action="/findtweets" method="post">
<div><textarea name="content" rows="3" cols="60"></textarea></div>
<div><input type="submit" value="Search"></div>
</form>
<table border="1">
<th>Topic</th>
<th>User</th>
<th>Buzz</th>""")
topics = Topic.all()
for buzz in topics:
self.response.out.write("""
<tr>
<td>%s</td>
<td>%s<br><img src="%s" /></td>
<td>%s</td>
</tr>""" % (cgi.escape(buzz.topic),
cgi.escape(buzz.user),
cgi.escape(buzz.photo_url),
cgi.escape(buzz.text)))
self.response.out.write("""
</table>
<form action="/findtweets" method="get">
<div><input type="submit" value="Tweet Contents"></div>
</form>
</body>
</html>""")
def post(self):
"""search tweets and populate db"""
keywords = cgi.escape(self.request.get('content'))
klist = [kw for kw in keywords.split('\n') if kw.strip()]
for keyword in klist:
tweetlist = searchTweets(keyword)
# the parent key (topic) of the tweet
par = Topic(key_name=keyword)
# We are using memcache to remember the buzz associated
# with each topic. (Assignment spec was ambiguous)
tbuzz = memcache.get('%s:buzz' % keyword)
if tbuzz is None:
tbuzz = getBuzz(tweetlist)
# Buzz-ed tweet expires in one minute (60 sec)
memcache.add('%s:buzz' % keyword, tbuzz, 60)
par.store(tbuzz)
for t in tweetlist:
tweet = Tweet(parent=par)
tweet.store(t)
self.redirect('/')
