for j in range(0,len(tmp_df.Time)-1):
flag=0
if tmp_df.Event_Type[j]=='KeyDown':
for k in range(j+1,len(tmp_df.Time)):
if tmp_df.Event_Type[k]=='KeyDown':
flag=1
chars=[tmp_df.Key_Code[j],tmp_df.Shift[j],tmp_df.Key_Code[k],tmp_df.Shift[k]]
latency=tmp_df.Time[k]-tmp_df.Time[j]
break
if flag==1 and latency<1.5:
feature=categorize(chars)
feature_list[0][feature]=feature_list[0][feature]+latency
feature_list[1][feature]=feature_list[1][feature]+1
mean_latencies=feature_list[0]/feature_list[1]
############################################################################
#Calculating hold times and number of backspaces and cpm
feature_list_holdtime=np.zeros((2,7))
backspaces=0
characters=0
cpm=float('NaN')
backspace_per_character=float('NaN')
for j in range(0,len(tmp_df.Time)-1):
conn = mongo.Connection()
db = conn.CompProb
tweets = db.tweets
trainer = db.trainer
words = db.words
priors = count_sentiment(trainer)
#SELF_DESTRUCT(tweets, trainer, words)
cursor = tweets.find()
i = 0
# # while i < 4:
for tweet in cursor:
i += 1
cat = categorize(tweet['text'].lower())
tweet['category'] = cat
# tweet['max_sentiment'] = determine_sentiment(tweet['sentiment_stats'])
tweets.save(tweet)
if i % 1000 == 0:
print i
# tweet = tweets.find_one()
# print tweet['text']
# text = split_tweet(tweet['text'])
# print text
# result = classify(text, priors, words)
# print result
# sentiment = determine_sentiment(result)
# print sentiment
# #i += 1
flag = 0
if tmp_df.Event_Type[j] == 'KeyDown':
for k in range(j + 1, len(tmp_df.Time)):
if tmp_df.Event_Type[k] == 'KeyDown':
flag = 1
chars = [
tmp_df.Key_Code[j], tmp_df.Shift[j],
tmp_df.Key_Code[k], tmp_df.Shift[k]
]
latency = tmp_df.Time[k] - tmp_df.Time[j]
break
if flag == 1 and latency < 1.5:
feature = categorize(chars)
feature_list[0][feature] = feature_list[0][feature] + latency
feature_list[1][feature] = feature_list[1][feature] + 1
mean_latencies = feature_list[0] / feature_list[1]
############################################################################
#Calculating hold times and number of backspaces and cpm
feature_list_holdtime = np.zeros((2, 7))
backspaces = 0
characters = 0
cpm = float('NaN')
backspace_per_character = float('NaN')
for j in range(0, len(tmp_df.Time) - 1):
db = conn.CompProb tweets = db.tweets trainer = db.trainer words = db.words priors = count_sentiment(trainer) #SELF_DESTRUCT(tweets, trainer, words) cursor = tweets.find() i = 0 # # while i < 4: for tweet in cursor: i += 1 cat = categorize(tweet['text'].lower()) tweet['category'] = cat # tweet['max_sentiment'] = determine_sentiment(tweet['sentiment_stats']) tweets.save(tweet) if i % 1000 == 0: print i # tweet = tweets.find_one() # print tweet['text'] # text = split_tweet(tweet['text']) # print text # result = classify(text, priors, words) # print result # sentiment = determine_sentiment(result) # print sentiment # #i += 1
