import gradio as gr
from transformers import pipeline
from langdetect import detect
import matplotlib.pyplot as plt

# Load multilingual sentiment model
sentiment_model = pipeline("sentiment-analysis", model="nlptown/bert-base-multilingual-uncased-sentiment")

# Emotion model (English-based)
emotion_model = pipeline("text-classification", model="j-hartmann/emotion-english-distilroberta-base", return_all_scores=True)

def analyze_text(text):
    try:
        # Detect language
        lang = detect(text)
    except:
        lang = "unknown"

    # Sentiment analysis
    sentiment_result = sentiment_model(text)[0]
    sentiment_label = sentiment_result['label']
    confidence = round(sentiment_result['score'], 4)

    # Emotion analysis
    emotions = emotion_model(text)[0]
    emotion_scores = {e["label"]: round(e["score"], 4) for e in emotions}

    # Plot emotion chart
    plt.figure(figsize=(6, 3))
    plt.bar(emotion_scores.keys(), emotion_scores.values(), color="skyblue")
    plt.title("Emotion Confidence Levels")
    plt.ylabel("Score")
    plt.xticks(rotation=45)
    plt.tight_layout()
    plt.savefig("emotion_chart.png")
    plt.close()

    return {
        "Detected Language": lang,
        "Sentiment": sentiment_label,
        "Confidence": confidence,
        "Emotions": emotion_scores
    }, "emotion_chart.png"

with gr.Blocks(title="K1ng Analyzer AI") as demo:
    gr.Image("K1nganalyzer_logo.png", elem_id="logo", show_label=False, height=120)
    gr.Markdown("### 🧠 Welcome to **K1ng Analyzer AI** — Smart Multilingual Emotion & Sentiment Analyzer 🌍")
    # (Your interface components go here)

# Gradio interface
demo = gr.Interface(
    fn=analyze_text,
    inputs=gr.Textbox(label="Enter text to analyze"),
    outputs=[
        gr.JSON(label="Analysis Result"),
        gr.Image(label="Emotion Confidence Chart")
    ],
    title="K1ng Analyzer V3 🌍🧠",
    description="Multilingual Sentiment + Emotion Analyzer with Visualization"
)

demo.launch()