import streamlit as st
import pandas as pd
import requests
import csv

# Define your API key and endpoint
api_key = 'AIzaSyAQ4dXlOkF8rPC21f6omTS4p6v-uJ2vVIg'
url = "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash-exp:generateContent"
headers = {'Content-Type': 'application/json'}

# Cache the sentiment analysis function to improve performance
@st.cache_data
def analyze_sentiment(text):
    """
    Analyze the sentiment of the given text using the Gemini API.
    """
    system_prompt = """
    You are a Sentiment Analysis Tool (SEA). Analyze the following comments and classify the sentiment of each as positive, neutral, or negative.
    Return the results in the following format:
    Comment: <comment>
    Sentiment: <sentiment>
    ---
    Additionally, provide actionable insights into customer satisfaction trends in the following format:
    ### Suggestions for Improvement:
    - <suggestion 1>
    - <suggestion 2>
    """
    data = {
        "contents": [{
            "parts": [{"text": f"{system_prompt}\n\n{text}"}]
        }]
    }
    response = requests.post(url, headers=headers, json=data, params={'key': api_key})
    if response.status_code == 200:
        return response.json()
    else:
        st.error(f"Request failed with status code {response.status_code}: {response.text}")
        return None

def process_file(file):
    """
    Process the uploaded file, analyze sentiment for all comments, and display results.
    """
    file_text = file.read().decode('utf-8')
    sentiment_result = analyze_sentiment(file_text)
    if sentiment_result:
        response_text = sentiment_result.get('candidates', [{}])[0].get('content', {}).get('parts', [{}])[0].get('text', '').strip()

        # Parse comments and sentiments
        results = []
        lines = response_text.split('\n')
        comment = None
        sentiment = None
        for line in lines:
            if line.startswith("Comment:"):
                comment = line.replace("Comment:", "").strip()
            elif line.startswith("Sentiment:"):
                sentiment = line.replace("Sentiment:", "").strip()
                if comment and sentiment:
                    results.append((comment, sentiment))
                    comment = None
                    sentiment = None

        # Display results
        df = pd.DataFrame(results, columns=['Comment/Prompt', 'Sentiment'])
        st.write("### Predicted Sentiments")
        st.dataframe(df)

        # Sentiment distribution
        sentiment_counts = df['Sentiment'].value_counts().reset_index()
        sentiment_counts.columns = ['Sentiment', 'Count']
        with st.expander("View Sentiment Distribution"):
            st.bar_chart(sentiment_counts.set_index('Sentiment'))

        # Suggestions
        suggestions = []
        current_section = None
        for line in lines:
            if line.startswith("### Suggestions for Improvement:"):
                current_section = "Suggestions for Improvement"
            elif current_section and line.startswith("- "):
                suggestions.append(line.replace("- ", "").strip())
        if suggestions:
            st.write("### Suggestions for Improvement")
            for suggestion in suggestions:
                st.write(f"- {suggestion}")
        else:
            st.warning("No suggestions available.")

        # CSV download
        output_file = 'sentiment_analysis_results.csv'
        df.to_csv(output_file, index=False)
        st.download_button(
            label="Download Results as CSV",
            data=open(output_file, 'rb').read(),
            file_name=output_file,
            mime='text/csv',
        )
    else:
        st.error("Sentiment analysis failed.")

# Streamlit layout
st.set_page_config(page_title="Sentiment Analysis Tool", layout="wide")
st.title("Sentiment Analysis Tool (SEA) 💬")
st.write("Analyze customer feedback with sentiment classification and actionable insights.")

# Sidebar for instructions
with st.sidebar:
    st.header("Instructions 📄")
    st.write("""
    1. Upload a CSV file containing customer feedback in the main area.
    2. Analyze real-time feedback using the text input box.
    3. Download sentiment analysis results as a CSV file.
    """)
    st.write("---")
    st.header("About")
    st.write("This app uses the Gemini API for sentiment analysis and provides actionable insights.")

# Main layout with tabs
tab1, tab2 = st.tabs(["📁 File Analysis", "✍️ Real-Time Feedback"])

with tab1:
    st.write("### Upload a CSV file for batch sentiment analysis:")
    uploaded_file = st.file_uploader("Choose a CSV file", type="csv")
    if uploaded_file is not None:
        with st.spinner("Processing file..."):
            process_file(uploaded_file)

with tab2:
    st.write("### Enter your feedback for real-time analysis:")
    feedback_input = st.text_area("Enter your feedback:", placeholder="Type your feedback here...")
    if st.button("Analyze Sentiment"):
        if feedback_input.strip() == "":
            st.warning("Please enter some feedback to analyze.")
        else:
            with st.spinner("Analyzing sentiment..."):
                sentiment_result = analyze_sentiment(feedback_input)
                if sentiment_result:
                    sentiment = sentiment_result.get('candidates', [{}])[0].get('content', {}).get('parts', [{}])[0].get('text', '').strip().lower()
                    if "positive" in sentiment:
                        st.success(f"Sentiment: **Positive** 😊")
                    elif "neutral" in sentiment:
                        st.info(f"Sentiment: **Neutral** 😐")
                    elif "negative" in sentiment:
                        st.error(f"Sentiment: **Negative** 😠")
                    else:
                        st.warning(f"Sentiment: **Unknown** 🤔")
                else:
                    st.error("Sentiment analysis failed.")