import gradio as gr
from transformers import pipeline
import pandas as pd
import matplotlib.pyplot as plt
from wordcloud import WordCloud
import plotly.express as px

# Initialize sentiment analysis pipeline
sentiment_pipeline = pipeline("sentiment-analysis", model='distilbert-base-uncased-finetuned-sst-2-english')

def categorize_sentiment(score):
    emoji_mapping = {
        'Highly Dissatisfied': '😡',
        'Dissatisfied': '😞',
        'Neutral': '😐',
        'Satisfied': '😊',
        'Highly Satisfied': '😄'
    }

    if score >= 0.8:
        return 'Highly Satisfied', emoji_mapping['Highly Satisfied']
    elif score >= 0.1:
        return 'Satisfied', emoji_mapping['Satisfied']
    elif score >= -0.1:
        return 'Neutral', emoji_mapping['Neutral']
    elif score >= -0.8:
        return 'Dissatisfied', emoji_mapping['Dissatisfied']
    else:
        return 'Highly Dissatisfied', emoji_mapping['Highly Dissatisfied']

def analyze_sentiment(input_text):
    result = sentiment_pipeline(input_text)[0]
    label = result['label']
    score = result['score']

    if label == "NEGATIVE":
        score = -score  # Make the score negative if the label is negative

    emoji = "😊" if label == "POSITIVE" else "😞"
    sentiment_label_with_emoji = f'{label} {emoji}'
    sentiment_category, emoji = categorize_sentiment(score)

    return str(score), sentiment_label_with_emoji, sentiment_category

def analyze_sentiments_from_excel(file):
    try:
        df = pd.read_excel(file.name)
    except Exception as e:
        return str(e), None, None, None, None  # Return error message if file cannot be read

    results = []

    for text in df['Text']:
        result = sentiment_pipeline(text)[0]
        label = result['label']
        score = result['score']

        if label == "POSITIVE":
            score = abs(score)
        else:
            score = -abs(score)

        sentiment_label, emoji = categorize_sentiment(score)
        results.append((text, score, label, sentiment_label, emoji))

    results_df = pd.DataFrame(results, columns=['Text', 'Sentiment Score', 'Sentiment Label', 'Sentiment Category', 'Emoji'])

    # Calculate percentages for each sentiment category
    sentiment_counts = results_df['Sentiment Category'].value_counts(normalize=True) * 100

    # Create pie chart for sentiment labels
    fig_labels = px.pie(results_df, names='Sentiment Label', title='Sentiment Label Distribution')

    # Create histogram for sentiment scores
    fig_scores = px.histogram(results_df, x='Sentiment Score', title='Sentiment Score Distribution')

    # Generate Word Cloud
    text = " ".join(results_df['Text'].tolist())
    wordcloud = WordCloud(width=800, height=400, random_state=21, max_font_size=110).generate(text)
    plt.figure(figsize=(12, 8))
    plt.imshow(wordcloud, interpolation="bilinear")
    plt.axis('off')
    plt.title('Word Cloud')
    plt.tight_layout()
    wordcloud_fig = plt

    return results_df, sentiment_counts, fig_labels, fig_scores, wordcloud_fig


def switch_app(app_name):
    if app_name == "Sentiment Analysis":
        iface1.show()
        iface2.hide()
    else:
        iface1.hide()
        iface2.show()

with gr.Blocks() as demo:
    with gr.Row():
        with gr.Column():
            input_text = gr.Textbox(lines=5, label="Input Text", placeholder="Kindly type in your product review")
            output_text = [
                gr.Textbox(label="Sentiment Score"),
                gr.Textbox(label="Sentiment Label"),
                gr.HTML(label="Sentiment Category")
            ]
            iface1 = gr.Interface(analyze_sentiment,
                                 inputs=input_text,
                                 outputs=output_text,
                                 title='<span style="color: yellow">Sentiment Analysis</span>',
                                 description='Explore the Emotions in Text. Experience the power of sentiment analysis with our tool. Get precise sentiment scores, emotive emojis, and categorized sentiments for any text.',
                                 theme='dark',
                                 css="""body {background-color: #121212; font-family: Arial, sans-serif;}
                                      .gradio {box-shadow: none; border-radius: 10px;}
                                      .input {background-color: #1E1E1E; color: #FFFFFF; border: none;}
                                      .output {background-color: #1E1E1E; color: #FFFFFF; border: none;}
                                      .output p {margin: 5px;}
                                      .output div {margin: 5px;}
                                      .output .highlight {padding: 5px; border-radius: 5px;}""")

        with gr.Column():
            file = gr.File(label="Upload Excel file")
            output_text = [
                gr.Dataframe(label="Text Analysis"),
                gr.Textbox(label="Sentiment Category"),
                gr.Plot(label="Sentiment Label"),
                gr.Plot(label="Sentiment Score"),
                gr.Plot(label="Word Cloud")
            ]
            iface2 = gr.Interface(analyze_sentiments_from_excel,
                                 inputs=file,
                                 outputs=output_text,
                                 title='<span style="color: yellow">Sentiment Analysis from Excel</span>',
                                 description='Upload an Excel file with sentiment texts and predict sentiment scores and labels. Analyze the sentiment distribution.',
                                 theme='dark',
                                 css="""body {background-color: #121212; font-family: Arial, sans-serif;}
                                      .gradio {box-shadow: none; border-radius: 10px;}
                                      .input {background-color: #1E1E1E; color: #FFFFFF; border: none;}
                                      .output {background-color: #1E1E1E; color: #FFFFFF; border: none;}
                                      .output p {margin: 5px;}
                                      .output div {margin: 5px;}
                                      .output .highlight {padding: 5px; border-radius: 5px;}""")

    with gr.Row():
        switch_button = gr.Button("Switch App")
        switch_button.click(switch_app, inputs=gr.Dropdown(["Sentiment Analysis", "Sentiment Analysis from Excel"], label="Select App"), outputs=None)

demo.launch()