Update app.py (#4)

- Update app.py (e836caaad6df9b25afabd777cfb75667141b0fe0)

app.py

@@ -1,308 +1,308 @@
-import PyPDF2
-import streamlit as st
-from transformers import pipeline
-import io
-from datetime import datetime, timedelta
-import gspread
-from google.oauth2.service_account import Credentials
-import pandas as pd
-import matplotlib.pyplot as plt
-from io import BytesIO
-from reportlab.lib.pagesizes import letter
-from reportlab.pdfgen import canvas
-import speech_recognition as sr  # Speech recognition package
-
-# Google Sheets setup
-SCOPE = ["https://spreadsheets.google.com/feeds", "https://www.googleapis.com/auth/drive"]
-CREDS_PATH = r"C:\Users\Muthuraja\Downloads\modern-cycling-444916-g6-82c207d3eb47.json"  # Provide your Google credentials path
-
-# Initialize the Hugging Face QA pipeline
-qa_pipeline = pipeline("question-answering")
-
-# Initialize Sentiment Analysis Pipeline using Hugging Face
-sentiment_model = pipeline('sentiment-analysis')
-
-# Initialize Google Sheets connection
-def initialize_google_sheets():
-    credentials = Credentials.from_service_account_file(CREDS_PATH, scopes=SCOPE)
-    try:
-        client = gspread.authorize(credentials)
-        sheet = client.open("SalesStores").sheet1  # Change Google Sheet name to "SalesStores"
-        return sheet
-    except gspread.exceptions.APIError as e:
-        st.error(f"Google Sheets API error: {e}")
-        return None
-
-sheet = initialize_google_sheets()
-
-# Function to extract text from PDF using PyPDF2
-def extract_pdf_text_with_pypdf(pdf_file):
-    pdf_text = ""
-    with io.BytesIO(pdf_file.read()) as pdf_data:
-        pdf_reader = PyPDF2.PdfReader(pdf_data)
-        for page_num in range(len(pdf_reader.pages)):
-            page = pdf_reader.pages[page_num]
-            pdf_text += page.extract_text()
-    return pdf_text.strip()
-
-# Function to answer a query using Hugging Face's QA pipeline
-def answer_query(question, context):
-    result = qa_pipeline(question=question, context=context)
-    return result['answer']
-
-# Function to analyze sentiment using Hugging Face's pre-trained model
-def analyze_sentiment(text):
-    sentiment = sentiment_model(text)[0]  # Output is a list of dictionaries
-    label = sentiment['label']
-    score = sentiment['score']
-    
-    # Define sentiment labels
-    if label == "POSITIVE" and score > 0.6:
-        sentiment_description = "Positive"
-    elif label == "NEGATIVE" and score < 0.4:  # Adjust threshold for negative sentiment
-        sentiment_description = "Negative"
-    else:
-        sentiment_description = "Neutral"
-    
-    return score, sentiment_description
-
-# Function to update Google Sheets without product name
-def update_sheet_without_product(sentiment_score, sentiment_description, relevant_answer):
-    if sheet:
-        timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-        sheet.append_row([timestamp, sentiment_description, sentiment_score, relevant_answer, "No Product Name"])
-    else:
-        st.error("Google Sheets connection not initialized.")
-
-# Function to suggest product recommendations based on sentiment or query
-def suggest_product_recommendations(sentiment_description, query):
-    # Based on the sentiment, recommend a product type
-    if "laptop" in query.lower():
-        if sentiment_description == "Positive":
-            recommendations = [
-                "Check out this powerful gaming laptop!",
-                "How about this lightweight ultrabook?",
-                "This laptop has amazing reviews, you should consider it!"
-            ]
-        elif sentiment_description == "Negative":
-            recommendations = [
-                "Maybe you'd prefer a different laptop brand?",
-                "Check out these budget-friendly laptops instead."
-            ]
-        else:
-            recommendations = [
-                "Looking for a laptop? Here's a variety of options for you!"
-            ]
-    elif "smartphone" in query.lower():
-        if sentiment_description == "Positive":
-            recommendations = [
-                "Check out our best-selling smartphone!",
-                "This smartphone has incredible features and great reviews."
-            ]
-        elif sentiment_description == "Negative":
-            recommendations = [
-                "Looking for an alternative smartphone? Here's something else you might like.",
-                "You might like these highly-rated budget smartphones instead."
-            ]
-        else:
-            recommendations = [
-                "Looking for the latest smartphone? Here's a range of options."
-            ]
-    elif "headphones" in query.lower():
-        if sentiment_description == "Positive":
-            recommendations = [
-                "These headphones are highly rated and perfect for music lovers!",
-                "Check out this noise-cancelling model."
-            ]
-        elif sentiment_description == "Negative":
-            recommendations = [
-                "Here are some alternatives that might suit your needs better.",
-                "Consider these wireless headphones for more comfort."
-            ]
-        else:
-            recommendations = [
-                "Here's a list of headphones with great sound quality!"
-            ]
-    else:
-        recommendations = [
-            "Based on your query, here are some great options across different categories!"
-        ]
-    
-    return recommendations
-
-# Function to filter data by date
-def filter_data_by_date(data, date_filter):
-    if date_filter == "Today":
-        start_date = datetime.now().replace(hour=0, minute=0, second=0, microsecond=0)
-        data = data[data['Timestamp'] >= start_date]
-    elif date_filter == "One Week":
-        start_date = datetime.now() - timedelta(weeks=1)
-        data = data[data['Timestamp'] >= start_date]
-    return data
-
-# Function to generate PDF for the call history
-def generate_pdf(data):
-    buffer = BytesIO()
-    c = canvas.Canvas(buffer, pagesize=letter)
-    c.setFont("Helvetica", 10)
-    y_position = 750
-    c.drawString(30, y_position, "Call History Report")
-    y_position -= 20
-    for index, row in data.iterrows():
-        c.drawString(30, y_position, f"Sentiment: {row['Sentiment']}, Answer: {row['Answer']}, Product: {row['Product Name']}")
-        y_position -= 15
-        if y_position <= 40:
-            c.showPage()
-            c.setFont("Helvetica", 10)
-            y_position = 750
-    c.save()
-    buffer.seek(0)
-    return buffer
-
-# Function to listen to speech and convert it to text
-def listen_to_speech():
-    recognizer = sr.Recognizer()
-    with sr.Microphone() as source:
-        recognizer.adjust_for_ambient_noise(source)  # Adjust for background noise
-        st.write("Listening...")  # Optional: Add a message to indicate listening state
-        try:
-            audio = recognizer.listen(source, timeout=5, phrase_time_limit=10)  # Listen for the audio input
-            st.write("Recognizing...")  # Optional: Add a message for recognition process
-            text = recognizer.recognize_google(audio)  # Use Google's speech recognition to convert audio to text
-            st.write(f"Recognized: {text}")
-            return text  # Return the text detected from the audio
-        except sr.UnknownValueError:
-            st.error("Sorry, I could not understand the audio.")  # Handle case when the audio is unclear
-            return None
-        except sr.RequestError:
-            st.error("Could not request results from Google Speech Recognition service.")  # Handle network issues
-            return None
-        except Exception as e:
-            st.error(f"An error occurred: {e}")
-            return None
-
-# Function to suggest related follow-up questions based on the answer
-def suggest_related_questions():
-    related_questions = [
-        "Can you explain more about the product?",
-        "What are the features of this product?",
-        "How does it compare to other products?",
-        "Can I get more details about the specifications?",
-        "What is the price of the product?"
-    ]
-    return related_questions
-
-# Dashboard functions
-def display_dashboard():
-    # Fetch data from Google Sheets
-    if sheet:
-        data = pd.DataFrame(sheet.get_all_records())  # Load all rows into a DataFrame
-
-        # Ensure the Timestamp column exists and is in datetime format
-        if 'Timestamp' in data.columns:
-            data['Timestamp'] = pd.to_datetime(data['Timestamp'])
-
-            # Add a date filter to the dashboard
-            date_filter = st.selectbox("Filter by Date", ["All Time", "Today", "One Week"])
-
-            # Filter data based on the selected date range
-            if date_filter != "All Time":
-                data = filter_data_by_date(data, date_filter)
-
-            # Ensure we have required columns
-            if 'Sentiment' in data.columns and 'Answer' in data.columns:
-                # Product Relevance Chart: Show sentiment distribution for each product
-                st.subheader("Product Relevance Chart")
-                product_sentiment = data.groupby(['Product Name', 'Sentiment']).size().unstack().fillna(0)
-
-                fig, ax = plt.subplots(figsize=(8, 6))
-                product_sentiment.plot(kind='bar', stacked=True, ax=ax, colormap='viridis')
-                ax.set_ylabel("Number of Queries")
-                ax.set_xlabel("Product Name")
-                ax.set_title("Sentiment Distribution for Each Product")
-                st.pyplot(fig)
-
-                # Plot sentiment distribution overall
-                sentiment_counts = data['Sentiment'].value_counts()
-                st.subheader("Overall Sentiment Distribution")
-                fig, ax = plt.subplots()
-                sentiment_counts.plot(kind='bar', ax=ax, color='lightcoral')
-                ax.set_ylabel("Frequency")
-                ax.set_xlabel("Sentiment")
-                st.pyplot(fig)
-
-                # Call Statistics
-                total_calls = len(data)
-                avg_sentiment = data['Sentiment'].apply(lambda x: 1 if x == 'Positive' else -1 if x == 'Negative' else 0).mean()
-                avg_sentiment = round(avg_sentiment, 2)
-
-                st.subheader("Call Activity Statistics")
-                st.write(f"Total Calls: {total_calls}")
-                st.write(f"Average Sentiment: {avg_sentiment}")
-
-                # Call History Table
-                st.subheader("Call History")
-                st.write(data[['Timestamp', 'Sentiment', 'Answer', 'Product Name']])
-
-                # Download option for the entire history (PDF)
-                pdf = generate_pdf(data)
-                st.download_button(
-                    label="Download Call History as PDF",
-                    data=pdf,
-                    file_name="call_history.pdf",
-                    mime="application/pdf"
-                )
-
-# Main Streamlit UI and workflow
-def main():
-    st.title('Real-Time Customer Query Analysis & Call History')
-
-    # Sidebar Navigation
-    sidebar_option = st.sidebar.selectbox("Select an Option", ["Dashboard", "Call Analysis"])
-
-    if sidebar_option == "Dashboard":
-        display_dashboard()
-
-    elif sidebar_option == "Call Analysis":
-        # Upload PDF file
-        uploaded_pdf = st.file_uploader("Upload a PDF file", type="pdf")
-        if uploaded_pdf:
-            pdf_text = extract_pdf_text_with_pypdf(uploaded_pdf)
-
-            if not pdf_text:
-                st.error("No text could be extracted from the PDF.")
-                return
-
-            # Speech recognition button
-            if st.button("Start Speech Recognition"):
-                user_input = listen_to_speech()
-                if user_input:
-                    # Sentiment Analysis
-                    sentiment_score, sentiment_description = analyze_sentiment(user_input)
-
-                    # Answer the query using the Hugging Face QA pipeline
-                    answer = answer_query(user_input, pdf_text)
-                    st.write(f"Answer: {answer}")
-
-                    # Display Sentiment Result
-                    st.write(f"Sentiment: {sentiment_description} (Score: {sentiment_score:.2f})")
-
-                    # Recommend products based on sentiment and query
-                    st.subheader("Recommended Products")
-                    recommendations = suggest_product_recommendations(sentiment_description, user_input)
-                    for rec in recommendations:
-                        st.write(f"- {rec}")
-
-                    # Store the query and the response in Google Sheets without the product name
-                    update_sheet_without_product(sentiment_score, sentiment_description, answer)
-
-         
-
-                    # Suggest related follow-up questions
-                    st.subheader("Related Follow-up Questions")
-                    related_questions = suggest_related_questions()
-                    for question in related_questions:
-                        st.write(f"- {question}")
-
-if __name__ == "__main__":
-    main()
+import PyPDF2
+import streamlit as st
+from transformers import pipeline
+import io
+from datetime import datetime, timedelta
+import gspread
+from google.oauth2.service_account import Credentials
+import pandas as pd
+import matplotlib.pyplot as plt
+from io import BytesIO
+from reportlab.lib.pagesizes import letter
+from reportlab.pdfgen import canvas
+import speech_recognition as sr  # Speech recognition package
+
+# Google Sheets setup
+SCOPE = ["https://spreadsheets.google.com/feeds", "https://www.googleapis.com/auth/drive"]
+CREDS_PATH = r"C:\Users\Muthuraja\OneDrive\Attachments\Desktop\second\modern-cycling-444916-g6-82c207d3eb47.json"  # Provide your Google credentials path
+
+# Initialize the Hugging Face QA pipeline
+qa_pipeline = pipeline("question-answering")
+
+# Initialize Sentiment Analysis Pipeline using Hugging Face
+sentiment_model = pipeline('sentiment-analysis')
+
+# Initialize Google Sheets connection
+def initialize_google_sheets():
+    credentials = Credentials.from_service_account_file(CREDS_PATH, scopes=SCOPE)
+    try:
+        client = gspread.authorize(credentials)
+        sheet = client.open("SalesStores").sheet1  # Change Google Sheet name to "SalesStores"
+        return sheet
+    except gspread.exceptions.APIError as e:
+        st.error(f"Google Sheets API error: {e}")
+        return None
+
+sheet = initialize_google_sheets()
+
+# Function to extract text from PDF using PyPDF2
+def extract_pdf_text_with_pypdf(pdf_file):
+    pdf_text = ""
+    with io.BytesIO(pdf_file.read()) as pdf_data:
+        pdf_reader = PyPDF2.PdfReader(pdf_data)
+        for page_num in range(len(pdf_reader.pages)):
+            page = pdf_reader.pages[page_num]
+            pdf_text += page.extract_text()
+    return pdf_text.strip()
+
+# Function to answer a query using Hugging Face's QA pipeline
+def answer_query(question, context):
+    result = qa_pipeline(question=question, context=context)
+    return result['answer']
+
+# Function to analyze sentiment using Hugging Face's pre-trained model
+def analyze_sentiment(text):
+    sentiment = sentiment_model(text)[0]  # Output is a list of dictionaries
+    label = sentiment['label']
+    score = sentiment['score']
+    
+    # Define sentiment labels
+    if label == "POSITIVE" and score > 0.6:
+        sentiment_description = "Positive"
+    elif label == "NEGATIVE" and score < 0.4:  # Adjust threshold for negative sentiment
+        sentiment_description = "Negative"
+    else:
+        sentiment_description = "Neutral"
+    
+    return score, sentiment_description
+
+# Function to update Google Sheets without product name
+def update_sheet_without_product(sentiment_score, sentiment_description, relevant_answer):
+    if sheet:
+        timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
+        sheet.append_row([timestamp, sentiment_description, sentiment_score, relevant_answer, "No Product Name"])
+    else:
+        st.error("Google Sheets connection not initialized.")
+
+# Function to suggest product recommendations based on sentiment or query
+def suggest_product_recommendations(sentiment_description, query):
+    # Based on the sentiment, recommend a product type
+    if "laptop" in query.lower():
+        if sentiment_description == "Positive":
+            recommendations = [
+                "Check out this powerful gaming laptop!",
+                "How about this lightweight ultrabook?",
+                "This laptop has amazing reviews, you should consider it!"
+            ]
+        elif sentiment_description == "Negative":
+            recommendations = [
+                "Maybe you'd prefer a different laptop brand?",
+                "Check out these budget-friendly laptops instead."
+            ]
+        else:
+            recommendations = [
+                "Looking for a laptop? Here's a variety of options for you!"
+            ]
+    elif "smartphone" in query.lower():
+        if sentiment_description == "Positive":
+            recommendations = [
+                "Check out our best-selling smartphone!",
+                "This smartphone has incredible features and great reviews."
+            ]
+        elif sentiment_description == "Negative":
+            recommendations = [
+                "Looking for an alternative smartphone? Here's something else you might like.",
+                "You might like these highly-rated budget smartphones instead."
+            ]
+        else:
+            recommendations = [
+                "Looking for the latest smartphone? Here's a range of options."
+            ]
+    elif "headphones" in query.lower():
+        if sentiment_description == "Positive":
+            recommendations = [
+                "These headphones are highly rated and perfect for music lovers!",
+                "Check out this noise-cancelling model."
+            ]
+        elif sentiment_description == "Negative":
+            recommendations = [
+                "Here are some alternatives that might suit your needs better.",
+                "Consider these wireless headphones for more comfort."
+            ]
+        else:
+            recommendations = [
+                "Here's a list of headphones with great sound quality!"
+            ]
+    else:
+        recommendations = [
+            "Based on your query, here are some great options across different categories!"
+        ]
+    
+    return recommendations
+
+# Function to filter data by date
+def filter_data_by_date(data, date_filter):
+    if date_filter == "Today":
+        start_date = datetime.now().replace(hour=0, minute=0, second=0, microsecond=0)
+        data = data[data['Timestamp'] >= start_date]
+    elif date_filter == "One Week":
+        start_date = datetime.now() - timedelta(weeks=1)
+        data = data[data['Timestamp'] >= start_date]
+    return data
+
+# Function to generate PDF for the call history
+def generate_pdf(data):
+    buffer = BytesIO()
+    c = canvas.Canvas(buffer, pagesize=letter)
+    c.setFont("Helvetica", 10)
+    y_position = 750
+    c.drawString(30, y_position, "Call History Report")
+    y_position -= 20
+    for index, row in data.iterrows():
+        c.drawString(30, y_position, f"Sentiment: {row['Sentiment']}, Answer: {row['Answer']}, Product: {row['Product Name']}")
+        y_position -= 15
+        if y_position <= 40:
+            c.showPage()
+            c.setFont("Helvetica", 10)
+            y_position = 750
+    c.save()
+    buffer.seek(0)
+    return buffer
+
+# Function to listen to speech and convert it to text
+def listen_to_speech():
+    recognizer = sr.Recognizer()
+    with sr.Microphone() as source:
+        recognizer.adjust_for_ambient_noise(source)  # Adjust for background noise
+        st.write("Listening...")  # Optional: Add a message to indicate listening state
+        try:
+            audio = recognizer.listen(source, timeout=5, phrase_time_limit=10)  # Listen for the audio input
+            st.write("Recognizing...")  # Optional: Add a message for recognition process
+            text = recognizer.recognize_google(audio)  # Use Google's speech recognition to convert audio to text
+            st.write(f"Recognized: {text}")
+            return text  # Return the text detected from the audio
+        except sr.UnknownValueError:
+            st.error("Sorry, I could not understand the audio.")  # Handle case when the audio is unclear
+            return None
+        except sr.RequestError:
+            st.error("Could not request results from Google Speech Recognition service.")  # Handle network issues
+            return None
+        except Exception as e:
+            st.error(f"An error occurred: {e}")
+            return None
+
+# Function to suggest related follow-up questions based on the answer
+def suggest_related_questions():
+    related_questions = [
+        "Can you explain more about the product?",
+        "What are the features of this product?",
+        "How does it compare to other products?",
+        "Can I get more details about the specifications?",
+        "What is the price of the product?"
+    ]
+    return related_questions
+
+# Dashboard functions
+def display_dashboard():
+    # Fetch data from Google Sheets
+    if sheet:
+        data = pd.DataFrame(sheet.get_all_records())  # Load all rows into a DataFrame
+
+        # Ensure the Timestamp column exists and is in datetime format
+        if 'Timestamp' in data.columns:
+            data['Timestamp'] = pd.to_datetime(data['Timestamp'])
+
+            # Add a date filter to the dashboard
+            date_filter = st.selectbox("Filter by Date", ["All Time", "Today", "One Week"])
+
+            # Filter data based on the selected date range
+            if date_filter != "All Time":
+                data = filter_data_by_date(data, date_filter)
+
+            # Ensure we have required columns
+            if 'Sentiment' in data.columns and 'Answer' in data.columns:
+                # Product Relevance Chart: Show sentiment distribution for each product
+                st.subheader("Product Relevance Chart")
+                product_sentiment = data.groupby(['Product Name', 'Sentiment']).size().unstack().fillna(0)
+
+                fig, ax = plt.subplots(figsize=(8, 6))
+                product_sentiment.plot(kind='bar', stacked=True, ax=ax, colormap='viridis')
+                ax.set_ylabel("Number of Queries")
+                ax.set_xlabel("Product Name")
+                ax.set_title("Sentiment Distribution for Each Product")
+                st.pyplot(fig)
+
+                # Plot sentiment distribution overall
+                sentiment_counts = data['Sentiment'].value_counts()
+                st.subheader("Overall Sentiment Distribution")
+                fig, ax = plt.subplots()
+                sentiment_counts.plot(kind='bar', ax=ax, color='lightcoral')
+                ax.set_ylabel("Frequency")
+                ax.set_xlabel("Sentiment")
+                st.pyplot(fig)
+
+                # Call Statistics
+                total_calls = len(data)
+                avg_sentiment = data['Sentiment'].apply(lambda x: 1 if x == 'Positive' else -1 if x == 'Negative' else 0).mean()
+                avg_sentiment = round(avg_sentiment, 2)
+
+                st.subheader("Call Activity Statistics")
+                st.write(f"Total Calls: {total_calls}")
+                st.write(f"Average Sentiment: {avg_sentiment}")
+
+                # Call History Table
+                st.subheader("Call History")
+                st.write(data[['Timestamp', 'Sentiment', 'Answer', 'Product Name']])
+
+                # Download option for the entire history (PDF)
+                pdf = generate_pdf(data)
+                st.download_button(
+                    label="Download Call History as PDF",
+                    data=pdf,
+                    file_name="call_history.pdf",
+                    mime="application/pdf"
+                )
+
+# Main Streamlit UI and workflow
+def main():
+    st.title('Real-Time Customer Query Analysis & Call History')
+
+    # Sidebar Navigation
+    sidebar_option = st.sidebar.selectbox("Select an Option", ["Dashboard", "Call Analysis"])
+
+    if sidebar_option == "Dashboard":
+        display_dashboard()
+
+    elif sidebar_option == "Call Analysis":
+        # Upload PDF file
+        uploaded_pdf = st.file_uploader("Upload a PDF file", type="pdf")
+        if uploaded_pdf:
+            pdf_text = extract_pdf_text_with_pypdf(uploaded_pdf)
+
+            if not pdf_text:
+                st.error("No text could be extracted from the PDF.")
+                return
+
+            # Speech recognition button
+            if st.button("Start Speech Recognition"):
+                user_input = listen_to_speech()
+                if user_input:
+                    # Sentiment Analysis
+                    sentiment_score, sentiment_description = analyze_sentiment(user_input)
+
+                    # Answer the query using the Hugging Face QA pipeline
+                    answer = answer_query(user_input, pdf_text)
+                    st.write(f"Answer: {answer}")
+
+                    # Display Sentiment Result
+                    st.write(f"Sentiment: {sentiment_description} (Score: {sentiment_score:.2f})")
+
+                    # Recommend products based on sentiment and query
+                    st.subheader("Recommended Products")
+                    recommendations = suggest_product_recommendations(sentiment_description, user_input)
+                    for rec in recommendations:
+                        st.write(f"- {rec}")
+
+                    # Store the query and the response in Google Sheets without the product name
+                    update_sheet_without_product(sentiment_score, sentiment_description, answer)
+
+         
+
+                    # Suggest related follow-up questions
+                    st.subheader("Related Follow-up Questions")
+                    related_questions = suggest_related_questions()
+                    for question in related_questions:
+                        st.write(f"- {question}")
+
+if __name__ == "__main__":
+    main()