In [150]:

    

#Takes Twitter archive CSV and grabs all source tweets


import csv
import redis
from markov.markov import add_line_to_index
from markov import Markov
import re

GRUBER = re.compile(ur'(?i)\b((?:https?://|www\d{0,3}[.]|[a-z0-9.\-]+[.][a-z]{2,4}/)(?:[^\s()<>]+|\(([^\s()<>]+|(\([^\s()<>]+\)))*\))+(?:\(([^\s()<>]+|(\([^\s()<>]+\)))*\)|[^\s`!()\[\]{};:\'".,<>?\xab\xbb\u201c\u201d\u2018\u2019]))')

riordan = []
with open ('tweets/tweets.csv', 'rb') as tweesv:
  tweetreader = csv.reader(tweesv, delimiter=',', quotechar='"')
  for row in tweetreader:
    if not row[6] and not row[5].startswith('RT @') and not row[5].startswith('rt @') and not row[5].startswith('@'):
      tweettext = row[5]
      for url in GRUBER.findall(tweettext):
        tweettext=tweettext.replace(url[0],'')
      
      riordan.append(tweettext)

#print riordan


client = redis.Redis()
twitter_data = Markov(prefix="twitter")
for tweet in riordan:
  add_line_to_index(tweet.split(), client, prefix="tweets")

tweet_data = Markov(prefix="tweets")


    

In [244]:

    

newTweet()


    

    




---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-244-193658413f9b> in <module>()
----> 1 newTweet()

NameError: name 'newTweet' is not defined

In [390]:

    

def newTweet():
    tweetline = tweet_data.generate()
    tweet = ""
    for word in tweetline:
        tweet += word + " "
    
    tweet = tweet.rstrip()
    
    if len(tweet)>140:
        return newTweet()
    else:
        return tweet


    

In [800]:

    

print newTweet()


    

    




completely re-imagined what a catalog can be. Totally ground-breaking #dpla

In [385]:

    




    

In [385]:

    




    

In [ ]: