def GET(self):
query = web.ctx.get('query')
html = html_helper.HTMLHelper()
twitterData = get_twitter_data.TwitterData()
if query:
if(query[0] == '?'):
query = query[1:]
arr = query.split('&')
logging.warning(arr)
#default values
time = 'daily'
for item in arr:
if 'keyword' in item:
keyword = item.split('=')[1]
elif 'method' in item:
method = item.split('=')[1]
elif 'time' in item:
time = item.split('=')[1]
#end loop
if(method != 'baseline' and method != 'naivebayes' and method != 'maxentropy' and method != 'svm'):
return html.getDefaultHTML(error=2)
tweets = twitterData.getTwitterData(keyword, time)
if(tweets):
if(method == 'baseline'):
bc = baseline_classifier.BaselineClassifier(tweets, keyword, time)
bc.classify()
return bc.getHTML()
elif(method == 'naivebayes'):
trainingDataFile = 'data/training_neatfile_2.csv'
#classifierDumpFile = 'data/naivebayes_trained_model.pickle'
classifierDumpFile = 'data/test/naivebayes_test_model.pickle'
trainingRequired = 0
nb = naive_bayes_classifier.NaiveBayesClassifier(tweets, keyword, time, \
trainingDataFile, classifierDumpFile, trainingRequired)
nb.classify()
return nb.getHTML()
elif(method == 'maxentropy'):
trainingDataFile = 'data/training_neatfile.csv'
classifierDumpFile = 'data/maxent_trained_model.pickle'
trainingRequired = 0
maxent = max_entropy_classifier.MaxEntClassifier(tweets, keyword, time, \
trainingDataFile, classifierDumpFile, trainingRequired)
maxent.classify()
return maxent.getHTML()
elif(method == 'svm'):
trainingDataFile = 'data/training_neatfile.csv'
classifierDumpFile = 'data/svm_trained_model.pickle'
trainingRequired = 0
sc = libsvm_classifier.SVMClassifier(tweets, keyword, time, \
trainingDataFile, classifierDumpFile, trainingRequired)
sc.classify()
return sc.getHTML()
else:
return html.getDefaultHTML(error=1)
else:
return html.getDefaultHTML()
def search():
keyword = request.args.get('keyword')
method = request.args.get('method')
time = 'daily'
time = request.args.get('time')
html = html_helper.HTMLHelper()
#print html
twitterData = get_twitter_data.TwitterData()
#print twitterData
if keyword:
if(method != 'baseline' and method != 'naivebayes' and method != 'maxentropy'):
return html.getDefaultHTML(error=2)
ACCESS_TOKEN = session['oauth_token']
ACCESS_TOKEN_SECRET = session['oauth_secret']
tweets = twitterData.getTwitterData(keyword, time,CONSUMER_KEY, CONSUMER_SECRET, ACCESS_TOKEN, ACCESS_TOKEN_SECRET)
print tweets,"-------------------------"
if(tweets):
if(method == 'baseline'):
bc = baseline_classifier.BaselineClassifier(tweets, keyword, time)
bc.classify()
return bc.getHTML()
elif(method == 'naivebayes'):
trainingDataFile = 'data/training_neatfile_2.csv'
classifierDumpFile = 'data/naivebayes_trained_model.pickle'
#classifierDumpFile = 'data/test/naivebayes_test_model.pickle'
trainingRequired = 0
nb = naive_bayes_classifier.NaiveBayesClassifier(tweets, keyword, time, \
trainingDataFile, classifierDumpFile, trainingRequired)
nb.classify()
return nb.getHTML()
elif(method == 'maxentropy'):
trainingDataFile = 'data/training_neatfile.csv'
classifierDumpFile = 'data/maxent_trained_model.pickle'
trainingRequired = 0
maxent = max_entropy_classifier.MaxEntClassifier(tweets, keyword, time, \
trainingDataFile, classifierDumpFile, trainingRequired)
maxent.classify()
return maxent.getHTML()
else:
return html.getDefaultHTML(error=1)
else:
return html.getDefaultHTML()
infile = open('data/tweets_current.pickle')
tweets = pickle.load(infile)
infile.close()
#algorithm = 'baseline'
#algorithm = 'naivebayes'
#algorithm = 'maxent'
#algorithm = 'svm'
if (len(sys.argv) < 2):
print "Please choose the algorithm to test, syntax = python analyze.py (svm|naivebayes|maxent)"
exit()
algorithm = sys.argv[1]
if (algorithm == 'baseline'):
bc = baseline_classifier.BaselineClassifier(tweets, keyword, time)
bc.classify()
val = bc.getHTML()
elif (algorithm == 'naivebayes'):
#trainingDataFile = 'data/training_trimmed.csv'
trainingDataFile = 'data/full_training_dataset.csv'
classifierDumpFile = 'data/test/naivebayes_test_model.pickle'
trainingRequired = 1
print 'Started to instantiate Naive Bayes Classifier'
sys.stdout.flush()
nb = naive_bayes_classifier.NaiveBayesClassifier(tweets, keyword, time,\
trainingDataFile, classifierDumpFile, trainingRequired)
#nb.classify()
print 'Computing Accuracy'
sys.stdout.flush()
nb.accuracy()
print(
"\n\tBaseline classifier has been already applied. Please Move to next steps :)\n"
)
else:
print("\tSystem exiting now. Visit again !!!\n")
sys.exit()
elif (inp == '1'):
if (os.path.isfile("sentiment_tweet_list.txt")):
print(
"\tBaseline classifier has been already applied. Please Move to next steps :)\n"
)
else:
print("\tworking on it\n")
df = pd.read_excel("Stemmed_DATA.xlsx", sheet_name="Sheet1")
list_tweets = df['Stemming tweet']
bc = baseline_classifier.BaselineClassifier(list_tweets)
bc.classify()
bc.writeOutput('sentiment_tweet_list.txt', 'w')
else:
print("\tWrong choice. System exiting now.\n")
sys.exit()
print("Split the data into Train and Test dataset.\n")
inp = input("Press 0/1 to continue ... ")
if (inp == '0'):
if (os.path.isfile("Data_to_split.xlsx")
and os.path.isfile("train_data.xlsx")
and os.path.isfile("test_data.xlsx")):
print("\tAll required files already exist. :)\n")
else:
print("\tSystem exiting now. Visit again and then press 1 !!!\n")