def fetch_posts(self, query: dict, start_date: datetime,
end_date: datetime) -> list:
"""Fetch tweets from archive.org that match the given query for the given day."""
ua = UserAgent()
ua.set_max_size(100 * 1024 * 1024)
ua.set_timeout(90)
ua.set_timing([1, 2, 4, 8, 16, 32, 64, 128, 256, 512])
end_date = end_date + datetime.timedelta(days=1)
start_arg = start_date.strftime('%Y-%m-%d')
end_arg = end_date.strftime('%Y-%m-%d')
enc_query = urlencode({
'q': query,
'date_from': start_arg,
'date_to': end_arg
})
url = "https://searchtweets.archivelab.org/export?" + enc_query
log.debug("archive.org url: " + url)
response = ua.get(url)
if not response.is_success():
raise McPostsArchiveTwitterDataException(
"error fetching posts: " + response.decoded_content())
decoded_content = response.decoded_content()
# sometimes we get null characters, which choke the csv module
decoded_content = decoded_content.replace('\x00', '')
meta_tweets = []
lines = decoded_content.splitlines()[1:]
for row in csv.reader(lines, delimiter="\t"):
fields = 'user_name user_screen_name lang text timestamp_ms url'.split(
' ')
meta_tweet = {}
for i, field in enumerate(fields):
meta_tweet[field] = row[i] if i < len(row) else ''
if 'url' not in meta_tweet or meta_tweet['url'] == '':
log.warning("meta_tweet '%s' does not have a url" % str(row))
continue
meta_tweet['tweet_id'] = get_tweet_id_from_url(meta_tweet['url'])
meta_tweets.append(meta_tweet)
add_tweets_to_meta_tweets(meta_tweets)
return meta_tweets
def fetch_posts_from_api(self, query: str, start_date: datetime,
end_date: datetime) -> list:
"""Fetch day of tweets from crimson hexagon and twitter."""
meta_tweets = []
next_cursor = None
while True:
data = self._fetch_posts_from_api_single_page(
query, start_date, end_date, next_cursor)
meta_tweets = meta_tweets + data['results']
if 'nextCursor' in data:
next_cursor = data['nextCursor']
else:
break
if 'results' not in data:
raise McPostsBWTwitterDataException("Unknown response status: " +
str(data))
for mt in meta_tweets:
try:
mt['tweet_id'] = get_tweet_id_from_url(mt['url'])
except McTwitterUrlException:
raise McPostsBWTwitterQueryException("""
Unable to parse tweet url %s. Make sure brandwatch query only includes twitter as a source.
""" % mt['url'])
add_tweets_to_meta_tweets(meta_tweets)
posts = []
for mt in meta_tweets:
log.debug("mt: %d" % mt['tweet_id'])
if 'tweet' in mt:
publish_date = dateutil.parser.parse(
mt['tweet']['created_at']).isoformat()
post = {
'post_id': str(mt['tweet_id']),
'data': mt,
'content': mt['tweet']['text'],
'publish_date': publish_date,
'author': mt['tweet']['user']['screen_name'],
'channel': mt['tweet']['user']['screen_name'],
'url': mt['url']
}
posts.append(post)
return posts
def fetch_posts_from_api(
self,
query: str,
start_date: datetime,
end_date: datetime,
sample: Optional[int] = None,
page_size: Optional[int] = None,
) -> list:
"""Fetch day of tweets from crimson hexagon and twitter."""
decoded_content = self._get_content_from_api(query, start_date,
end_date)
assert sample is None, "Sampling is not implemented."
assert page_size is None, "Page size limiting is not supported."
data = dict(decode_json(decoded_content))
if 'status' not in data or not data['status'] == 'success':
raise McPostsCHTwitterDataException("Unknown response status: " +
str(data))
meta_tweets = data['posts']
for mt in meta_tweets:
mt['tweet_id'] = get_tweet_id_from_url(mt['url'])
add_tweets_to_meta_tweets(meta_tweets)
posts = []
for mt in meta_tweets:
log.debug("mt: %d" % mt['tweet_id'])
if 'tweet' in mt:
publish_date = dateutil.parser.parse(
mt['tweet']['created_at']).isoformat()
post = {
'post_id': str(mt['tweet_id']),
'data': mt,
'content': mt['tweet']['text'],
'publish_date': publish_date,
'author': mt['tweet']['user']['screen_name'],
'channel': mt['tweet']['user']['screen_name'],
'url': mt['url']
}
posts.append(post)
return posts
def fetch_posts_from_api(
self,
query: str,
start_date: datetime,
end_date: datetime,
sample: Optional[int] = None,
page_size: Optional[int] = None,
) -> list:
"""Fetch day of tweets."""
if page_size is None:
page_size = 5000
meta_tweets = []
next_cursor = None
while True:
data = self._fetch_posts_from_api_single_page(
query=query,
start_date=start_date,
end_date=end_date,
next_cursor=next_cursor,
page_size=page_size,
)
meta_tweets = meta_tweets + data['results']
log.debug(f"Sample: {sample}; meta_tweets: {len(meta_tweets)}")
if 'nextCursor' not in data or (sample is not None and len(meta_tweets) >= sample):
break
else:
next_cursor = data['nextCursor']
if 'results' not in data:
raise McPostsBWTwitterDataException("Unknown response status: " + str(data))
for mt in meta_tweets:
try:
mt['tweet_id'] = get_tweet_id_from_url(mt['url'])
except McTwitterUrlException:
raise McPostsBWTwitterQueryException(
"""
Unable to parse tweet url %s. Make sure brandwatch query only includes twitter as a source.
""" % mt['url'])
add_tweets_to_meta_tweets(meta_tweets)
posts = []
for mt in meta_tweets:
log.debug("mt: %d" % mt['tweet_id'])
if 'tweet' in mt:
publish_date = dateutil.parser.parse(mt['tweet']['created_at']).isoformat()
post = {
'post_id': str(mt['tweet_id']),
'data': mt,
'content': mt['tweet']['text'],
'publish_date': publish_date,
'author': mt['tweet']['user']['screen_name'],
'channel': mt['tweet']['user']['screen_name'],
'url': mt['url']
}
posts.append(post)
return posts
def fetch_posts(self, query: str, start_date: datetime,
end_date: datetime) -> list:
"""Fetch day of tweets from crimson hexagon"""
ch_monitor_id = int(query)
log.debug("crimson_hexagon_twitter.fetch_posts")
ua = UserAgent()
ua.set_max_size(100 * 1024 * 1024)
ua.set_timeout(90)
ua.set_timing([1, 2, 4, 8, 16, 32, 64, 128, 256, 512])
config = TopicsMineConfig()
api_key = config.crimson_hexagon_api_key()
end_date = end_date + datetime.timedelta(days=1)
start_arg = start_date.strftime('%Y-%m-%d')
end_arg = end_date.strftime('%Y-%m-%d')
url = (
"https://api.crimsonhexagon.com/api/monitor/posts?auth=%s&id=%d&start=%s&end=%s&extendLimit=true"
% (api_key, ch_monitor_id, start_arg, end_arg))
log.debug("crimson hexagon url: " + url)
response = ua.get(url)
if not response.is_success():
raise McPostsCHTwitterDataException("error fetching posts: " +
response.decoded_content())
decoded_content = response.decoded_content()
data = dict(decode_json(decoded_content))
if 'status' not in data or not data['status'] == 'success':
raise McPostsCHTwitterDataException("Unknown response status: " +
str(data))
meta_tweets = data['posts']
for mt in meta_tweets:
mt['tweet_id'] = get_tweet_id_from_url(mt['url'])
add_tweets_to_meta_tweets(meta_tweets)
posts = []
for mt in meta_tweets:
log.warning("mt: %d" % mt['tweet_id'])
if 'tweet' in mt:
post = {
'post_id': mt['tweet_id'],
'data': mt,
'content': mt['tweet']['text'],
'publish_date': mt['tweet']['created_at'],
'author': mt['tweet']['user']['screen_name'],
'channel': mt['tweet']['user']['screen_name'],
'url': mt['url']
}
posts.append(post)
return posts
