mathdatech/twoot
1
0
Fork 0
mirror of https://gitlab.com/jeancf/twoot.git synced 2025-03-27 07:56:48 +00:00
Code Issues 1 Packages Projects Releases Wiki Activity

Extracted tweet_id and status_id

Browse Source
This commit is contained in:
jeancf 2020-12-16 21:55:13 +01:00
parent c25e36b498
commit 7cc076053f
1 changed files with 5 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
twoot.py
View File

@ -235,12 +235,11 @@ def main(argv):
# Extract twitter timeline # Extract twitter timeline
timeline = soup.find_all('div', class_='timeline-item') timeline = soup.find_all('div', class_='timeline-item')
print(len(timeline))
exit(0)
logging.info('Processing timeline')
for status in timeline: for status in timeline:
# Extract tweet ID and status ID # Extract tweet ID and status ID
tweet_id = str(status['href']).strip('?p=v') tweet_id = status.find('a', class_='tweet-link').get('href').strip('#m')
status_id = tweet_id.split('/')[3] status_id = tweet_id.split('/')[3]
logging.debug('processing tweet %s', tweet_id) logging.debug('processing tweet %s', tweet_id)
@ -250,10 +249,6 @@ def main(argv):
(twit_account, mast_instance, mast_account, tweet_id)) (twit_account, mast_instance, mast_account, tweet_id))
tweet_in_db = db.fetchone() tweet_in_db = db.fetchone()
logging.debug("SELECT * FROM toots WHERE twitter_account='{}' AND mastodon_instance='{}' AND mastodon_account='{}' AND tweet_id='{}'"
.format(twit_account, mast_instance, mast_account, tweet_id)
)
if tweet_in_db is not None: if tweet_in_db is not None:
logging.debug("Tweet %s already in database", tweet_id) logging.debug("Tweet %s already in database", tweet_id)
# Skip to next tweet # Skip to next tweet
@ -262,13 +257,12 @@ def main(argv):
logging.debug('Tweet %s not found in database', tweet_id) logging.debug('Tweet %s not found in database', tweet_id)
reply_to_username = None reply_to_username = None
# Check if the tweet is a reply-to # TODO Check if the tweet is a reply-to
reply_to_div = status.find('div', class_='tweet-reply-context username') reply_to_div = None
if reply_to_div is not None: if reply_to_div is not None:
# Do we need to handle reply-to tweets? # Do we need to handle reply-to tweets?
if tweets_and_replies: if tweets_and_replies:
# Capture user name being replied to # TODO Capture user name being replied to
reply_to_username = reply_to_div.a.get_text()
else: else:
# Skip this tweet # Skip this tweet
logging.debug("Tweet is a reply-to and we don't want that. Skipping.") logging.debug("Tweet is a reply-to and we don't want that. Skipping.")