def getValue():
filter = request.form['filter']
keyword = str(filter)
obj = StreamingClass(keyword,10)
obj.findTweets()
# Tweets data
tweetIDList,tweetUserNameList,tweetTextList = obj.getDetails()
#####
# Covert into JSON data format
dataList = []
for i in range(0,len(tweetIDList)):
row = {}
row["id"] = tweetIDList[i]
row["userName"] = tweetUserNameList[i]
row["text"] = tweetTextList[i]
try:
row["sentiment"] = gcpNLP(tweetTextList[i])
except:
row["sentiment"] = gcpNLP(langTranslator(tweetTextList[i]))
# dataDictionary[str(i)] = row
dataList.append(row)
return render_template('index.html',analysis=dataList)
def main():
testTweet = "@libertlady @TheGreySonof @sergiolpn @EugeniaRolon_ @Angelalerena Corrección; Foxconn tiene denuncias por trabajo e… https://t.co/XAuqim69GT"
cleanedTestTweet = cleanTweet(testTweet)
print(cleanedTestTweet)
# Using Google Cloud Platform NLP model
testTweetSentiment = gcpNLP(testTweet)
cleanedTestTweetSentiment = gcpNLP(cleanedTestTweet)
# Printing the results
print("testTweetSentiment", testTweetSentiment)
print("cleanedTestTweetSentiment", cleanedTestTweetSentiment)
def addSentimentProperties(listOfTuples):
# Add sentiment score and sentiment magnitude into the tuples
output = [] # Output list
for i in listOfTuples:
sentiment = gcpNLP(i[-1])
sentimentScore = sentiment.score
sentimentMagnitude = sentiment.magnitude
i = i + (sentimentScore, sentimentMagnitude
) # Add sentiment score and magnitude into the tuple
output.append(i)
return output
def main():
fileName = "TrainingPhrases.xlsx"
readObj = ReadExcelFile(fileName)
# Reading the phrases column
listOfPhrases = readObj.getColumnValues("Phrase")
# Passing the phrases into Google and getting sentiment parameters and storing new values into excel sheet
n = len(listOfPhrases)
# Creating an object to write values in Excel file
writeObj = WriteExcelFile(fileName)
for i in range(0, n):
sentimentScore, sentimentMagnitude = gcpNLP(listOfPhrases[i])
# For sentiment score
writeObj.write(i + 2, 3, str(sentimentScore))
# For sentiment magnitude
writeObj.write(i + 2, 4, str(sentimentMagnitude))
def getTextFromDatabase(dbName, tableName):
# Connect with Clickhouse
dbObject = ClickhouseClient("10.0.0.30")
# Get all table data
data = dbObject.selectData(dbName, tableName)
# Add new columns to existing data table (For sentiment score and magnitude)
# Need to be edited
dbObject.addColumn(dbObject, tableName, "score", "String")
dbObject.addColumn(dbObject, tableName, "magnitude", "String")
# Need to be edited
# Read the rows one by one
for i in data:
# Get the sentiment value one by one
sentiment = gcpNLP(i[-1])
print(sentiment.score)
def storeClickhouse(ipAddress, userName, tweetURL, listOfTweetReplyObjects):
# Creating Clickhouse client object
dbObject = ClickhouseClient(ipAddress)
# Create a new table
tableName = createTableName(userName, tweetURL)
# Creating list of tuples
listOfTuples = []
# Read tweet objects one by one and store them in Clickhouse
for i in listOfTweetReplyObjects:
# Filter out required data from the object
dateTime, userID, userName, textID, text = filterDataFields(i)
# Save date into tuple
dataTuple = (dateTime, userID, userName, textID, text)
# Pass the data to Google NLP API and fetch sentiment parameters
sentiment = gcpNLP(text)
sentimentScore = sentiment.score
sentimentMagnitude = sentiment.magnitude
# Append sentiment properties into dataTuple
dataTuple = dataTuple + (sentimentScore, sentimentMagnitude)
# Append to listOfTuples
listOfTuples.append(dataTuple)
# Convert list of tuples into list of dict
listOfDict = getListOfDict([
'date_time', 'user_id', 'user_name', 'text_id', 'text', 'score',
'magnitude'
], listOfTuples)
# Store data inside the Clickhouse
dbObject.createTable("twitter", tableName)
# Store the data
dbObject.insertData("twitter", tableName, listOfDict)
def modelTesting(phrase):
val1, val2 = gcpNLP(phrase)
# Parameters
W1 = np.array([[
0.7944864, 2.5306618, -4.5559416, 0.3835472, 0.99691343, -16.251022,
-4.3798785, 8.443347, -2.4882212, 0.18412943
],
[
1.3711807, 2.76434, -2.2028754, 1.7946438, 0.1288933,
-13.720011, -4.2971883, -8.429026, -2.3004491,
0.82452744
]])
W2 = np.array([[1.545527], [2.8403213], [4.8268833], [-2.4048],
[-1.0374035], [-13.8223915], [-3.170226], [12.88139],
[3.4203565], [1.0138328]])
b1 = np.array([[
-3.2464652e+00, -4.2906141e+00, 5.2482719e-03, -1.4363248e+00,
8.6730689e-01, 5.2704124e+00, 4.4298167e+00, 5.3570614e+00,
-1.5046034e-02, -2.4429324e+00
]])
b2 = np.array([[0.959692]])
# Inputs
inputArray = np.array([[val1, val2]])
# Layer one output
out1 = sigmoid(inputArray.dot(np.array(W1)) + np.array(b1))
# Layer two output
out2 = sigmoid(out1.dot(np.array(W2)) + np.array(b2))
return np.round(out2, 1)[0][0]