twitools/csvdb.py

#!/usr/bin/env python3

import tools

import sqlite3, csv, sys

def makeDB(path=tools.dbpath()):
 try:
  infile = open('tweets.csv')
 except IOError:
  raise IOError("Please make sure that the tweets.csv from the Twitter download is located in this directory.")
 
 input = list(csv.reader(infile))
 
 conn = sqlite3.connect(path)
 cur = conn.cursor()
 
 try:
  cur.execute("CREATE TABLE tweets(`tweet_id` INTEGER NOT NULL, `in_reply_to_status_id` TEXT, `in_reply_to_user_id` TEXT, `timestamp` TEXT, `source` TEXT, `text` TEXT, `retweeted_status_id` TEXT, `retweeted_status_user_id` TEXT, `retweeted_status_timestamp` TEXT, `expanded_urls` TEXT, PRIMARY KEY(tweet_id));")
 except sqlite3.OperationalError:
  raise IOError("%s already exists. Please delete it before trying to create a new one." % path)

 for row in input[1:]:
  cur.execute("INSERT INTO tweets VALUES(" + row[0].replace("'","''") + ",'" + row[1].replace("'","''")  + "','" + row[2].replace("'","''")  + "','" + row[3].replace("'","''")  + "','" + row[4].replace("'","''")  + "','" + row[5].replace("'","''")  + "','" + row[6].replace("'","''")  + "','" + row[7].replace("'","''")  + "','" + row[8].replace("'","''")  + "','" + row[9].replace("'","''")  + "');")
 
 conn.commit()

if __name__ == "__main__":
 if len(sys.argv) > 2:
  raise ValueError(sys.argv[0] + " only takes one argument, the path of the new database file.")
 try:
  makeDB(sys.argv[1])
 except IndexError:
  makeDB()