Ife branch

.streamlit/config.toml

@@ -0,0 +1,6 @@
+[theme]
+primaryColor="#c3112a"
+backgroundColor="#e0e1e6"
+secondaryBackgroundColor="#2abace"
+textColor="#121fe4"
+font="monospace"

base_app.py

@@ -21,62 +21,313 @@
 	https://docs.streamlit.io/en/latest/
 
 """
+
 # Streamlit dependencies
 import streamlit as st
 import joblib,os
+from PIL import Image
+import io
+from streamlit_option_menu import option_menu
+
 
 # Data dependencies
 import pandas as pd
+from nltk.corpus import stopwords
+from nltk.stem import WordNetLemmatizer
+import re
+import contractions
+import matplotlib.pyplot as plt
+#import seaborn as sns
 
 # Vectorizer
-news_vectorizer = open("resources/tfidfvect.pkl","rb")
+news_vectorizer = open("resources/vectoriser-ngram-(1,2).pickle","rb")
 tweet_cv = joblib.load(news_vectorizer) # loading your vectorizer from the pkl file
 
 # Load your raw data
 raw = pd.read_csv("resources/train.csv")
+##define function to preprocess data
+def preprocess(textdata):
+    processedTweet = []
+
+    # Create Lemmatizer and Stemmer.
+    wordLemm = WordNetLemmatizer()
+
+    # Defining regex patterns.
+    urlPattern        = r"http\S+"
+    punctuations      = "[^a-zA-Z#@_]"
+    sequencePattern   = r"(.)\1\1+"
+    seqReplacePattern = r"\1\1"
+
+    for tweet in textdata:
+        tweet = tweet.lower()
+
+        # Replace all URls with 'URL'
+        tweet = re.sub(urlPattern,'newsurl',tweet)
+        #expand contractions
+        tweet= contractions.fix(tweet)              
+        # Replace all punctuation.
+        tweet = re.sub(punctuations, " ", tweet)
+        # Replace 3 or more consecutive letters by 2 letter.
+        tweet = re.sub(sequencePattern, seqReplacePattern, tweet)
+
+        tweetwords = ''
+        for word in tweet.split():
+            # Checking if the word is a stopword.
+            #if word not in stopwordlist:
+            if word not in stopwords.words('english'):
+                if len(word)>3:
+                # Lemmatizing the word.
+                    word = wordLemm.lemmatize(word)
+                    tweetwords += (word+' ')
+
+        processedTweet.append(tweetwords)
+
+
+        return   processedTweet
+#define function to export csv file
+def convert_df(df):
+   	return df.to_csv().encode('utf-8')
+def predict(path, model, feature):
+	predictor = joblib.load(open(os.path.join(path,model),"rb"))
+	prediction = predictor.predict(feature)
+	return prediction
+#define function to format graph
+def my_fmt(x):
+    return '{:.4f}%\n({:.0f})'.format(x, total*x/100)
+
+logo = Image.open("resources/imgs/default_edited.png")
+oluyemi_new= Image.open('resources/imgs/yemi edsa picture.JPG')
+oluyemi=oluyemi_new.resize((160,250))
+joshua_new= Image.open('resources/imgs/Eujosh_pic_new.JPEG')
+joshua=joshua_new.resize((160,250))
+abiola_new= Image.open('resources/imgs/Abiola_pic.JPEG')
+abiola=abiola_new.resize((160,250))
+ifeoluwa_new= Image.open('resources/imgs/ifeoluwa_pic_new.JPEG')
+ifeoluwa=ifeoluwa_new.resize((160,250))
+lawson_new=Image.open('resources/imgs/Lawson_pic.JPEG')
+lawson=lawson_new.resize((160,250))
+stephen_new=Image.open('resources/imgs/Stephen_pic_new.JPEG')
+stephen=stephen_new.resize((160,250))
+welcome_message='<p style="font-family:Mono space; color:black; font-size: 20px;">Welcome, we are glad to have you here.\
+			Kindly use Navigation on the side to find your way around. Enjoy your stay</p>'
+purpose='<p style="font-family:Mono space; color:black; font-size: 15px; ">The Purpose of Our Organization</p>'
 
 # The main function where we will build the actual app
 def main():
 	"""Tweet Classifier App with Streamlit """
 
 	# Creates a main title and subheader on your page -
 	# these are static across all pages
-	st.title("Tweet Classifer")
-	st.subheader("Climate change tweet classification")
+	#st.title ("AIORIGIN")
+
+
 
 	# Creating sidebar with selection box -
 	# you can create multiple pages this way
-	options = ["Prediction", "Information"]
-	selection = st.sidebar.selectbox("Choose Option", options)
+	Navigation = ["About Us", "Classifier", "Information"]
+	selection = st.sidebar.selectbox("Navigation path", Navigation)
+
+	#Building the "About us" page
+	if selection== "About Us":
+		st.image(logo)
+		st.markdown(welcome_message,unsafe_allow_html=True)
+		with st.expander("Read about us"):
+			st.markdown(""" At AIORIGIN we make use of the wealth of knowledge of experts from different fields to \
+				provide cutting-edge solutions for businesses of all sizes.\
+					We analyse each situation and select the most appropriate innovative tools and \
+						proprietary technologies for the job at hand. \
+							The end result is a set of machine learning models and solutions tailored to the specific\
+								 needs and objectives of each of our clients.""")
+			st. markdown(purpose, unsafe_allow_html=True)
+			st.write("Our goal is to help our clients become more competitive and achieve results \
+					they coulld only have imagined. Innovative and proprietary development technologies,\
+						 exceptional services, and top-notch professional expertise are the means we \
+							use to achieve these objectives")
+		st.info("Meet the Team")
+		col1, col2, col3 = st.columns(3) #create images side by side
+		ife = Image.open("resources/imgs/ifeoluwa_pic_new.JPEG")
+		ife =ife.resize((160,250))
+		col1.write("Ifeoluwa Adeoti")
+		col1.image(ife)
+		col1.caption("Ife is the founder and CEO.")
+
+		yemi = Image.open('resources/imgs/yemi edsa picture.JPG')
+		yemi= yemi.resize((160,250))
+		col2.write("Oluyemi Alabi")
+		col2.image(yemi)
+		col2.caption("Yemi is the Chief Technical Officer.")
+
+		abiola = Image.open("resources/imgs/Abiola_pic.JPEG")
+		abiola=abiola.resize((160,250))
+		col3.write("Abiola Akinwale")
+		col3.image(abiola)
+		col3.caption("Abiola is the Creative Director.")
+
+		col4, col5, col6 = st.columns(3) #create images side by side
+
+		joshua = Image.open('resources/imgs/Eujosh_pic_new.JPEG')
+		joshua= joshua.resize((160,250))
+		col4.write("Joshua Umukoro")
+		col4.image(joshua)
+		col4.caption("Joshua is the Financial Head")
+
+		lawson = Image.open('resources/imgs/Lawson_pic.JPEG')
+		lawson= lawson.resize((160,250))
+		col5.write("Lawson Iduku")
+		col5.image(lawson)
+		col5.caption("Lawson is the Head of Human Resources")
+
+		stephen = Image.open('resources/imgs/Stephen_pic_new.JPEG')
+		stephen= stephen.resize((160,250))
+		col6.write("Stephen Tshiani ")
+		col6.image(stephen)
+		col6.caption("Stephen is the Sales Manager")
+
 
 	# Building out the "Information" page
 	if selection == "Information":
-		st.info("General Information")
+		st.info("Classifier Information")
 		# You can read a markdown file from supporting resources folder
-		st.markdown("Some information here")
-
+		st.markdown("The models below were initially used in training our data.")
+		st.write("1. Logistic Regression model")
+		st.write("2. Support Vector Classifier (SVC) model")
+		st.write("3. Naive-Bayes Model-Multinomial (MNB)")
+		st.write("4. K-Nearest Neighbour model (KNN)")
+		st.write("5. Random Forest model")
+		st.markdown(" Accuracy metric used is the F1_Score which takes into account of Precision and Recall score")
+		st.markdown(" Select *View model Performance graph* to see how they perform")
 		st.subheader("Raw Twitter data and label")
-		if st.checkbox('Show raw data'): # data is hidden if box is unchecked
+		if st.checkbox('Show sample of raw data'): # data is hidden if box is unchecked
 			st.write(raw[['sentiment', 'message']]) # will write the df to the page
+		elif st.checkbox('View model Perfomance graph'):
+			st.image("resources\imgs\chart.PNG")
+			with st.expander("See explanation"):
+				st.markdown("""The chart above shows model performance comparism. The chart on the left\
+					shows the F1_score for each model with SVC, Logistic Regression and Multinomial \
+						Naive-Bayes performing the best, while the chart on the right shows the comparism between training time, with SVC \
+							taking the longest time """)
+
 
 	# Building out the predication page
-	if selection == "Prediction":
-		st.info("Prediction with ML Models")
+	if selection == "Classifier":
+		st.title("Tweet Classifer")
+		st.subheader("Climate change tweet classification")
+		model_selection = st.radio("Please choose a model", ["Logistic Regression" ,"Multinomial Naive Bayes (Recommended)",\
+			"Linear Support Vector Classifier"], help= 'Select a model that will be used for prediction')
+		#graph_selection = st.radio("Please choose preferred graph", ["Bar chart" ,"Pie chart"], help= 'Select type of graph output')
+
+		st.info("NOTE: If you would like to analyse large amount of tweets at a go, kindly use the upload button on the side.  \
+			File to be uploaded must be '.csv' and have two columns. First column \
+				should be named 'index' while second column named 'tweets'. Maximum number of rows is '15,000' ")
+		st.info("For single tweets, enter the tweet to be analysed in the text area below")
 		# Creating a text box for user input
-		tweet_text = st.text_area("Enter Text","Type Here")
+		tweet_text = st.text_area("Enter Tweet","Type Here")
+
 
 		if st.button("Classify"):
+			#tweet_text= preprocess(tweet_text)
 			# Transforming user input with vectorizer
 			vect_text = tweet_cv.transform([tweet_text]).toarray()
+			#predictor = joblib.load(open(os.path.join("resources/model_5.pkl"),"rb"))
+			#prediction = predictor.predict(vect_text)
+
+			if model_selection == "Logistic Regression":				
+				# Load your .pkl file with the model of your choice + make predictions
+				# Try loading in multiple models to give the user a choice
+				prediction= predict('resources','Sentiment-LR.pickle', vect_text)
+				output = {-1: "an Anti-climate tweet", 0: "a Neutral tweet", 1: "a Pro-climate tweet", 2: "Climate change News"}
+				f = output.get(prediction[0])	
+			# When model has successfully run, will print prediction
+				st.success("Tweet is Categorized as: {}".format(f))
+
+			elif model_selection == "Multinomial Naive Bayes (Recommended)":
+				prediction=predict('resources','mnb.pickle', vect_text)
+				output = {-1: "an Anti-climate tweet", 0: "a Neutral tweet", 1: "a Pro-climate tweet", 2: "Climate change News"}
+				f = output.get(prediction[0])	
+			# When model has successfully run, will print prediction
+				st.success("Tweet is Categorized as: {}".format(f))
+
+			elif model_selection == "Linear Support Vector Classifier":
+				prediction=predict('resources','svc_model.pickle', vect_text)
+				output = {-1: "an Anti-climate tweet", 0: "a Neutral tweet", 1: "a Pro-climate tweet", 2: "Climate change News"}
+				f = output.get(prediction[0])	
+			# When model has successfully run, will print prediction
+				st.success("Tweet is Categorized as: {}".format(f))
+
 			# Load your .pkl file with the model of your choice + make predictions
 			# Try loading in multiple models to give the user a choice
-			predictor = joblib.load(open(os.path.join("resources/Logistic_regression.pkl"),"rb"))
-			prediction = predictor.predict(vect_text)
-
+			#output = {-1: "an Anti-climate tweet", 0: "a Neutral tweet", 1: "a Pro-climate tweet", 2: "Climate change News"}
+			#f = output.get(prediction[0])	
 			# When model has successfully run, will print prediction
-			# You can use a dictionary or similar structure to make this output
-			# more human interpretable.
-			st.success("Text Categorized as: {}".format(prediction))
+			#st.success("Tweet is Categorized as: {}".format(f))
+
+			## for the large test file, option to upload the file
+		upload_file= st.sidebar.file_uploader(
+			label="Upload the csv file containing tweets here",
+    		type="csv",
+   			accept_multiple_files=False,
+    		help='''Upload a csv file that contains tweet.     
+        	required structure:     
+        	first column = index;        
+        	second column = tweets;      
+        	first row = column headers;     
+        	length = max. 15,000 rows
+        	''')   
+		if upload_file is not None:
+			tweet_df = pd.read_csv(upload_file)
+			preprocess(tweet_df['tweets'])
+			vect_df = tweet_cv.transform(tweet_df['tweets']).toarray()
+
+			if model_selection =="Multinomial Naive Bayes (Recommended)":
+				#tweet_df = pd.read_csv(upload_file)
+				#vect_df = tweet_cv.transform(tweet_df['tweets']).toarray()
+				classification=predict('resources','mnb.pickle', vect_df)
+
+
+				#predicted = joblib.load(open(os.path.join("resources/mnb.pickle"),"rb"))
+				#classification = predicted.predict(vect_df)
+
+			elif model_selection == "Logistic Regression":
+				#vect_df = tweet_cv.transform(tweet_df['tweets']).toarray()
+				classification=predict('resources','Sentiment-LR.pickle', vect_df)
+				#predicted = joblib.load(open(os.path.join("resources/Sentiment-LR.pickle"),"rb"))
+				#classification = predicted.predict(vect_df)
+
+			elif model_selection == "Linear Support Vector Classifier":
+				classification=predict('resources','svc_model.pickle', vect_df)
+				#tweet_df = pd.read_csv(upload_file)
+				#vect_df = tweet_cv.transform(tweet_df['tweets']).toarray()
+					#classification=predict('resources','svc_model.pickle', vect_df)
+				#predicted = joblib.load(open(os.path.join("resources/svc_model.pickle"),"rb"))
+
+				#classification = predicted.predict(vect_df)
+
+
+			result = pd.DataFrame(classification, columns = ['sentiment'])
+			result['tweets'] = tweet_df['tweets']
+			result = result[['tweets', 'sentiment']]
+			m={'sentiment':{1:'Pro-climate change', 2: 'Climate change news', 
+			0:'Neutral', -1: 'Anti-climate change'}}
+			result.replace(m, inplace=True)
+			classified=convert_df(result)
+			st.download_button("Press to Download result file",
+			classified,"file.csv","text/csv",
+			key='download-csv')
+
+			graph=result['sentiment'].value_counts().plot(kind='pie', annotate=True)
+			graph= 'pie.png'
+			img = io.BytesIO()
+			plt.savefig(img, format='png')
+
+			st.download_button(
+				label="Download chart",
+				data=img,
+        		file_name=graph,
+        		mime="image/png")
+
+
+
+
 
 # Required to let Streamlit instantiate our web app.  
 if __name__ == '__main__':