from textblob import TextBlob
import gradio as gr
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt

def final_sentiment_app(text):
    if not text.strip():
        return "⚠️ Please enter some text!", None

    sentences = [s.strip() for s in text.replace('!','!|').replace('?','?|').replace('.','.|').split('|') if s.strip()]

    if len(sentences) == 0:
        return "⚠️ No sentences found!", None

    scores = []
    colors = []
    sentence_results = []
    pos_count = neg_count = neu_count = 0
    total_score = 0

    for i, sentence in enumerate(sentences):
        blob = TextBlob(sentence)
        score = round(blob.sentiment.polarity, 2)
        subjectivity = round(blob.sentiment.subjectivity, 2)
        scores.append(score)
        total_score += score

        if score >= 0.5:
            label = "🤩 Very Positive"
            colors.append('#00e676')
            pos_count += 1
        elif score > 0.1:
            label = "😊 Positive"
            colors.append('#00c853')
            pos_count += 1
        elif score <= -0.5:
            label = "😡 Very Negative"
            colors.append('#ff1744')
            neg_count += 1
        elif score < -0.1:
            label = "😞 Negative"
            colors.append('#d50000')
            neg_count += 1
        else:
            label = "😐 Neutral"
            colors.append('#aa00ff')
            neu_count += 1

        sentence_results.append(
            f"S{i+1} {label} | Score: {score} | Subjectivity: {subjectivity}\n"
            f"    → \"{sentence}\""
        )

    avg = round(total_score / len(sentences), 2)

    if avg >= 0.5:    overall = "🤩 Very Positive"
    elif avg > 0.1:   overall = "😊 Positive"
    elif avg <= -0.5: overall = "😡 Very Negative"
    elif avg < -0.1:  overall = "😞 Negative"
    else:             overall = "😐 Neutral"

    fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(11, 4))
    fig.patch.set_facecolor('#0d0d1a')
    fig.suptitle('Sentiment Analysis Report', color='white', fontsize=14, fontweight='bold')

    ax1.set_facecolor('#1a1a2e')
    x = range(len(sentences))
    bars = ax1.bar(x, scores, color=colors, edgecolor='white', linewidth=0.5, width=0.5)
    ax1.axhline(y=0, color='white', linewidth=1, linestyle='--', alpha=0.5)
    ax1.set_title('Score Per Sentence', color='white', fontsize=11)
    ax1.set_xticks(x)
    ax1.set_xticklabels([f'S{i+1}' for i in x], color='white')
    ax1.set_ylabel('Polarity (-1 to +1)', color='white')
    ax1.set_ylim(-1.2, 1.2)
    ax1.tick_params(colors='white')
    for bar, score in zip(bars, scores):
        ypos = bar.get_height()+0.05 if score >= 0 else bar.get_height()-0.12
        ax1.text(bar.get_x()+bar.get_width()/2, ypos,
                str(score), ha='center', color='white', fontsize=9, fontweight='bold')

    ax2.set_facecolor('#1a1a2e')
    pie_data = [(pos_count,'Positive\n','#00c853'),
                (neg_count,'Negative\n','#ff1744'),
                (neu_count,'Neutral\n','#aa00ff')]
    pie_data = [(v,l,c) for v,l,c in pie_data if v > 0]
    ax2.pie(
        [d[0] for d in pie_data],
        labels=[f"{d[1]}({d[0]})" for d in pie_data],
        colors=[d[2] for d in pie_data],
        autopct='%1.0f%%',
        textprops={'color':'white','fontsize':10},
        wedgeprops={'edgecolor':'white','linewidth':1.2},
        startangle=90
    )
    ax2.set_title('Sentiment Distribution', color='white', fontsize=11)
    plt.tight_layout()

    report = "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n"
    report += "📋 SENTENCE BREAKDOWN\n"
    report += "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n"
    report += "\n".join(sentence_results)
    report += "\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n"
    report += "📊 FINAL SUMMARY\n"
    report += "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n"
    report += f"🎯 Overall Sentiment  : {overall}\n"
    report += f"📈 Average Score      : {avg}\n"
    report += f"✅ Positive sentences : {pos_count}\n"
    report += f"❌ Negative sentences : {neg_count}\n"
    report += f"➖ Neutral sentences  : {neu_count}\n"
    report += f"📝 Total sentences    : {len(sentences)}\n"
    report += f"📏 Total words        : {len(text.split())}\n"
    report += "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"

    return report, fig

app = gr.Interface(
    fn=final_sentiment_app,
    inputs=gr.Textbox(
        placeholder="Paste a paragraph, product review, tweet or any text...",
        label="📝 Enter Your Text",
        lines=6
    ),
    outputs=[
        gr.Textbox(label="📋 Analysis Report", lines=18),
        gr.Plot(label="📈 Visual Charts")
    ],
    title="🧠 Sentiment Analyzer Pro",
    description="✨ AI-powered sentiment detection | Sentence-by-sentence breakdown | Visual charts | Built with Python & NLP",
    examples=[
        ["I love this college! The canteen food is terrible. Classes are okay. Teachers are amazing!"],
        ["This phone is fantastic! Battery life is poor. Camera quality is outstanding. Delivery was late."],
        ["I am so happy today! Work was stressful. But my friends made it better. Overall a good day!"],
        ["The movie was boring. Acting was terrible. But the music was absolutely amazing!"]
    ],
    theme="soft"
)

app.launch()