def gen_json_for_tweets_of_interest(data, identity_list):
of_id, uid_list = data
json_of_name = os.path.join(JSON_OUTPUT_DIRECTORY, str(of_id) + ".json.gz")
print 'inp: ', json_of_name, len(uid_list), uid_list[0:2]
tweets_to_write = []
if not os.path.exists(json_of_name):
for i, uid in enumerate(uid_list):
if i % 25 == 0:
print i, len(tweets_to_write)
try:
u = TwitterUser()
u.populate_tweets_from_file(os.path.join(
JSON_INPUT_DIRECTORY, uid + ".json.gz"),
store_json=True)
tweets_to_keep = []
for t in u.tweets:
if not t.retweeted and len(t.tokens) > 4:
expanded_token_set = copy(t.tokens)
for token in t.tokens:
expanded_token_set += get_alternate_wordforms(
token)
if len(set(expanded_token_set) & identity_list):
tweets_to_keep.append(t)
tweets_to_write += tweets_to_keep
except:
print 'FAILED JSON FOR USER: '******'WRITING JSON'
out_fil = gzip.open(json_of_name, "wb")
for tweet in tweets_to_write:
out_fil.write(
json.dumps(tweet.raw_json).strip().encode("utf8") + "\n")
out_fil.close()
def gen_conll_file(fil,ptb_dir, dp_dir):
user = TwitterUser()
user.populate_tweets_from_file(fil, do_tokenize=False)
if 50 <= user.n_total_tweets <= 15000 and\
user.followers_count <= 25000 and user.creation_date <= MIN_ACCOUNT_AGE:
dp_filename = os.path.join(dp_dir,str(user.user_id)+".gz")
ptb_filename = os.path.join(ptb_dir,str(user.user_id)+".txt.gz")
if not os.path.exists(dp_filename) or not os.path.exists(ptb_filename):
return ['no_dp_ptb',[user.user_id,os.path.exists(dp_filename),os.path.exists(ptb_filename)]]
penntreebank = {x[0] : x[1:] for x in read_grouped_by_newline_file(ptb_filename)}
dependency_parse = read_grouped_by_newline_file(dp_filename)
tweet_set = [(i,t) for i,t in enumerate(user.tweets) if t.retweeted is None and\
len(t.urls) == 0 and 'http:' not in t.text and\
langid.classify(t.text)[0] == 'en']
# non english speaker or spam
if len(tweet_set) < 40:
return ['notweets',user.user_id]
data_to_return = []
for twit_it, tweet in tweet_set:
data_for_tweet = []
ptb_for_tweet = penntreebank[str(tweet.id)]
dp_for_tweet = dependency_parse[twit_it]
if ptb_for_tweet[0].split("\t")[2] != DependencyParseObject(dp_for_tweet[0]).text:
print 'ahhhhh, weird stuff'
continue
for i, p in enumerate(dp_for_tweet):
d = DependencyParseObject(tsn([p,tweet.id,user.user_id,tweet.created_at.strftime("%m-%d-%y")],newline=False))
# get java features
spl_java = ptb_for_tweet[i].split("\t")
java_id, penn_pos_tag,word = spl_java[:3]
java_features = '' if len(spl_java) == 3 else spl_java[3]
d.features += [x for x in java_features.split("|") if x != '']
d.features.append("penn_treebank_pos="+penn_pos_tag)
data_for_tweet.append(d)
data_to_return.append(data_for_tweet)
return ['success', [user.user_id,data_to_return]]
else:
return ['baduser',user.user_id]
def gen_output(data, json_data_dir):
term, is_reply, tweets_needed = data
dataset = []
# get all user files
files = glob.glob(os.path.join(json_data_dir, "*"))
random.shuffle(files)
for f in files:
user = TwitterUser()
user.populate_tweets_from_file(f,
store_json=True,
do_arabic_stemming=False,
lemmatize=False)
if 50 <= user.n_total_tweets <= 10000 and\
user.followers_count <= 25000 and user.creation_date <= MIN_ACCOUNT_AGE:
tweet_set = [t for t in user.tweets if t.retweeted is None and\
len(t.urls) == 0 and 'http:' not in t.text and\
len(t.tokens) > 5 and\
t.created_at >= MIN_TWEET_DATE and\
(term == '' or term in t.tokens) and\
langid.classify(t.text)[0] == 'en'and\
sentiment(t.text)['compound'] != 0]
if is_reply:
tweet_set = [t for t in tweet_set if t.reply_to]
else:
tweet_set = [t for t in tweet_set if not t.reply_to]
if len(tweet_set) == 0:
print 'size 0', term, tweets_needed, is_reply
continue
tweet = random.sample(tweet_set, 1)[0]
print user.screen_name, term, tweets_needed, is_reply, ":::: ", tweet.text
dataset.append(tweet)
tweets_needed -= 1
if tweets_needed == 0:
name = term if term != '' else 'random'
name += '_reply' if is_reply else '_non_reply'
pickle.dump(dataset, open(name + ".p", 'wb'))
print 'done with: ', name, is_reply
return
else:
print 'failed user'
def gen_json_for_tweets_of_interest(input_filename,
output_filename,
keep_only_tweets_with_terms=None):
"""
This function generates a cleaned json file so that the identity
extraction only happens on "interesting" tweets. Right now,
interesting is defined as non-retweets that have >4 tokens. Feel free to redefine
as you feel is suitable
:param input_filename: input json file name (Can be gzipped)
:param output_filename: cleaned output json filename
:param keep_only_tweets_with_terms: If you only want to keep tweets containing a specific
set of terms, you can use this argument and pass in a set of terms here
:return:
"""
tweets_to_write = []
if not os.path.exists(output_filename):
try:
u = TwitterUser()
u.populate_tweets_from_file(input_filename, store_json=True)
tweets = [
t for t in u.tweets if not t.retweeted and len(t.tokens) > 4
]
tweets_to_keep = []
if keep_only_tweets_with_terms:
expanded_token_set = copy(t.tokens)
for t in tweets:
for token in t.tokens:
expanded_token_set += get_alternate_wordforms(token)
if len(
set(expanded_token_set)
& keep_only_tweets_with_terms):
tweets_to_keep.append(t)
else:
tweets_to_keep = tweets
tweets_to_write += tweets_to_keep
except:
print 'FAILED TO PARSE JSON FILE: ', input_filename
print 'WRITING JSON'
out_fil = gzip.open(output_filename, "wb")
for tweet in tweets_to_write:
out_fil.write(
json.dumps(tweet.raw_json).strip().encode("utf8") + "\n")
out_fil.close()
def gen_output(data, json_data_dir):
term,is_reply,tweets_needed = data
dataset = []
# get all user files
files = glob.glob(os.path.join(json_data_dir,"*"))
random.shuffle(files)
for f in files:
user = TwitterUser()
user.populate_tweets_from_file(f,store_json=True,do_arabic_stemming=False,lemmatize=False)
if 50 <= user.n_total_tweets <= 10000 and\
user.followers_count <= 25000 and user.creation_date <= MIN_ACCOUNT_AGE:
tweet_set = [t for t in user.tweets if t.retweeted is None and\
len(t.urls) == 0 and 'http:' not in t.text and\
len(t.tokens) > 5 and\
t.created_at >= MIN_TWEET_DATE and\
(term == '' or term in t.tokens) and\
langid.classify(t.text)[0] == 'en'and\
sentiment(t.text)['compound'] != 0]
if is_reply:
tweet_set = [t for t in tweet_set if t.reply_to]
else:
tweet_set = [t for t in tweet_set if not t.reply_to]
if len(tweet_set) == 0:
print 'size 0', term, tweets_needed, is_reply
continue
tweet = random.sample(tweet_set, 1)[0]
print user.screen_name, term, tweets_needed, is_reply, ":::: ", tweet.text
dataset.append(tweet)
tweets_needed -= 1
if tweets_needed == 0:
name = term if term != '' else 'random'
name += '_reply' if is_reply else '_non_reply'
pickle.dump(dataset,open(name+".p",'wb'))
print 'done with: ',name, is_reply
return
else:
print 'failed user'
def gen_json_for_tweets_of_interest(input_filename, output_filename,keep_only_tweets_with_terms=None):
"""
This function generates a cleaned json file so that the identity
extraction only happens on "interesting" tweets. Right now,
interesting is defined as non-retweets that have >4 tokens. Feel free to redefine
as you feel is suitable
:param input_filename: input json file name (Can be gzipped)
:param output_filename: cleaned output json filename
:param keep_only_tweets_with_terms: If you only want to keep tweets containing a specific
set of terms, you can use this argument and pass in a set of terms here
:return:
"""
tweets_to_write = []
if not os.path.exists(output_filename):
try:
u = TwitterUser()
u.populate_tweets_from_file(input_filename,store_json=True)
tweets = [t for t in u.tweets if not t.retweeted and len(t.tokens) > 4]
tweets_to_keep = []
if keep_only_tweets_with_terms:
expanded_token_set = copy(t.tokens)
for t in tweets:
for token in t.tokens:
expanded_token_set += get_alternate_wordforms(token)
if len(set(expanded_token_set) & keep_only_tweets_with_terms):
tweets_to_keep.append(t)
else:
tweets_to_keep = tweets
tweets_to_write += tweets_to_keep
except:
print 'FAILED TO PARSE JSON FILE: ', input_filename
print 'WRITING JSON'
out_fil = gzip.open(output_filename, "wb")
for tweet in tweets_to_write:
out_fil.write(json.dumps(tweet.raw_json).strip().encode("utf8") + "\n")
out_fil.close()
def getTweets(twitterid):
'''
Function to get the twitter data for an individual twitter ID.
This function is written to work with Kenny's github example here: https://github.com/kennyjoseph/twitter_dm
Input: string of twitterID
Output: list of the raw string of all tweets for twitterID
'''
from twitter_dm.TwitterUser import TwitterUser
tweets = []
u = TwitterUser()
u.populate_tweets_from_file(
twitterid + '.json'
) #Need to figure out of we can use numeric ID (123456789.json) or name (kenny_joseph.json)
for t in u.tweets:
tweets.append(
t.tokens
) #not sure if tokens is exactly what we want, we want the raw words, not necessarily tokens. We'll check this.
#
# texts={}
# source_filename='Datasets/Twitter/members.zip'
# parser = etree.XMLParser(encoding='utf8',recover=True)
# with zipfile.ZipFile(source_filename) as zf:
# for i,member in enumerate(zf.infolist()):
# name=member.filename.split('/')[1].split('.')[0] #filename is Raw3/name.csv
# if idx ==name:
# #print idx, name
# raw=zf.open(member)
# data=csv.reader(raw)
# for j,line in enumerate(data):
# if j>0:
# texts[idx+'_'+str(j)]=line[0]
# if texts=={}:
# print 'no tweets for ', idx
return tweets
from twitter_dm.TwitterUser import TwitterUser
import datetime
from time import mktime
u = TwitterUser()
u.populate_tweets_from_file("/Users/kennyjoseph/git/thesis/twitter_dm/examples/2431225676.json.gz")
for t in u.tweets:
print mktime(t.created_at.timetuple())
def gen_conll_file(fil, ptb_dir, dp_dir):
user = TwitterUser()
user.populate_tweets_from_file(fil, do_tokenize=False)
if 50 <= user.n_total_tweets <= 15000 and\
user.followers_count <= 25000 and user.creation_date <= MIN_ACCOUNT_AGE:
dp_filename = os.path.join(dp_dir, str(user.user_id) + ".gz")
ptb_filename = os.path.join(ptb_dir, str(user.user_id) + ".txt.gz")
if not os.path.exists(dp_filename) or not os.path.exists(ptb_filename):
return [
'no_dp_ptb',
[
user.user_id,
os.path.exists(dp_filename),
os.path.exists(ptb_filename)
]
]
penntreebank = {
x[0]: x[1:]
for x in read_grouped_by_newline_file(ptb_filename)
}
dependency_parse = read_grouped_by_newline_file(dp_filename)
tweet_set = [(i,t) for i,t in enumerate(user.tweets) if t.retweeted is None and\
len(t.urls) == 0 and 'http:' not in t.text and\
langid.classify(t.text)[0] == 'en']
# non english speaker or spam
if len(tweet_set) < 40:
return ['notweets', user.user_id]
data_to_return = []
for twit_it, tweet in tweet_set:
data_for_tweet = []
ptb_for_tweet = penntreebank[str(tweet.id)]
dp_for_tweet = dependency_parse[twit_it]
if ptb_for_tweet[0].split("\t")[2] != DependencyParseObject(
dp_for_tweet[0]).text:
print 'ahhhhh, weird stuff'
continue
for i, p in enumerate(dp_for_tweet):
d = DependencyParseObject(
tsn([
p, tweet.id, user.user_id,
tweet.created_at.strftime("%m-%d-%y")
],
newline=False))
# get java features
spl_java = ptb_for_tweet[i].split("\t")
java_id, penn_pos_tag, word = spl_java[:3]
java_features = '' if len(spl_java) == 3 else spl_java[3]
d.features += [x for x in java_features.split("|") if x != '']
d.features.append("penn_treebank_pos=" + penn_pos_tag)
data_for_tweet.append(d)
data_to_return.append(data_for_tweet)
return ['success', [user.user_id, data_to_return]]
else:
return ['baduser', user.user_id]
