Skip to content

Instantly share code, notes, and snippets.

@MihaiTabara
Last active September 8, 2021 16:03
Show Gist options
  • Save MihaiTabara/631ecb98f93046a9a454 to your computer and use it in GitHub Desktop.
Save MihaiTabara/631ecb98f93046a9a454 to your computer and use it in GitHub Desktop.
Script to download Twitter timeline for a user and store it to MongoDB
# script to download up to <= 3200 (the official API limit) of most recent tweets from a user's timeline
from pymongo import MongoClient
import tweepy
import json
#Twitter API credentials
CONSUMER_KEY = ''
CONSUMER_SECRET = ''
ACCESS_TOKEN = ''
ACCESS_TOKEN_SECRET = ''
class TwitterHarvester(object):
"""Create a new TwitterHarvester instance"""
def __init__(self, consumer_key, consumer_secret,
access_token, access_token_secret,
wait_on_rate_limit=False,
wait_on_rate_limit_notify=False):
self.auth = tweepy.OAuthHandler(consumer_key, consumer_secret)
self.auth.secure = True
self.auth.set_access_token(access_token, access_token_secret)
self.__api = tweepy.API(self.auth,
wait_on_rate_limit=wait_on_rate_limit,
wait_on_rate_limit_notify=wait_on_rate_limit_notify)
@property
def api(self):
return self.__api
def twitter_logic():
# instantiate an object of TwitterHarvester to use it's api object
# make sure to set the corresponding flags as True to whether or
# not automatically wait for rate limits to replenish
a = TwitterHarvester(CONSUMER_KEY, CONSUMER_SECRET,
ACCESS_TOKEN, ACCESS_TOKEN_SECRET,
wait_on_rate_limit=True,
wait_on_rate_limit_notify=True)
api = a.api
# assume there's MongoDB running on the machine, get a connection to it
conn = MongoClient('localhost', 27017)
db = conn['twitter_db']
collection = db['tweets']
# use the cursor to skip the handling of the pagination mechanism
# http://docs.tweepy.org/en/latest/cursor_tutorial.html
tweets = tweepy.Cursor(api.user_timeline, screen_name="cnnbrk").items()
while True:
# as long as I still have a tweet to grab
try:
data = tweets.next()
except StopIteration:
break
# convert from Python dict-like structure to JSON format
jsoned_data = json.dumps(data._json)
tweet = json.loads(jsoned_data)
# insert the information in the database
collection.insert(tweet)
if __name__ == "__main__":
twitter_logic()
@moltenform
Copy link

Thanks, this was very helpful.

Three suggestions: adding tweet_mode='extended' as a parameter to tweepy.Cursor lets you get the full, 280 character tweet instead of being truncated at 140. And it might be helpful to mention the parameter exclude_replies = True. Finally, one can wrap the tweepy.Cursor in limit_handled as described here, so that it pauses if it ever hits the access limit.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment