import gradio as gr
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

# ====================== MODEL ======================
model_path = "./best_model"
tokenizer = AutoTokenizer.from_pretrained(model_path)
model = AutoModelForSequenceClassification.from_pretrained(model_path)

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)
model.eval()

def predict_sentiment(text):
    if not text or text.strip() == "":
        return "Please enter text", "0.00"
    
    inputs = tokenizer(
        text,
        return_tensors="pt",
        truncation=True,
        padding=True,
        max_length=512
    ).to(device)
    
    with torch.no_grad():
        outputs = model(**inputs)
        probs = torch.softmax(outputs.logits, dim=1)
        pred_id = torch.argmax(probs, dim=1).item()
        confidence = torch.max(probs).item()
        label = model.config.id2label[pred_id]
    
    return label, f"{confidence:.2%}"

# ====================== PROFESSIONAL CSS ======================
custom_css = """
@import url('https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700;800&display=swap');

:root {
    --bg: #070A12;
    --panel: rgba(255, 255, 255, 0.04);
    --panel-strong: rgba(255, 255, 255, 0.07);
    --border: rgba(255, 255, 255, 0.08);
    --primary: #7C3AED;
    --secondary: #22D3EE;
    --text: #E5E7EB;
    --muted: #9CA3AF;
}

body {
    background: radial-gradient(circle at top, #0B1020, #05060A);
    font-family: 'Inter', sans-serif;
}

.gradio-container {
    max-width: 100% !important;
    margin: auto !important;
    padding: 40px 20px;
}

/* Main Card */
.block {
    background: var(--panel) !important;
    border: 1px solid var(--border) !important;
    border-radius: 24px !important;
    padding: 40px !important;
    backdrop-filter: blur(12px);
    box-shadow: 0 20px 60px rgba(0,0,0,0.6);
}

/* Title */
h1 {
    font-size: 46px !important;
    font-weight: 800 !important;
    text-align: center;
    color: white;
    letter-spacing: -1px;
}
h1 span {
    background: linear-gradient(90deg, #7C3AED, #22D3EE);
    -webkit-background-clip: text;
    -webkit-text-fill-color: transparent;
}

/* Subtitle */
.subtitle {
    text-align: center;
    color: var(--muted);
    font-size: 17px;
    margin-bottom: 30px;
}

/* Button */
button {
    background: linear-gradient(135deg, var(--primary), var(--secondary)) !important;
    color: white !important;
    font-weight: 700 !important;
    font-size: 17px !important;
    padding: 16px 0 !important;
    border-radius: 16px !important;
    width: 100%;
    margin: 15px 0 25px 0;
    box-shadow: 0 10px 30px rgba(124, 58, 237, 0.3);
}
button:hover {
    transform: translateY(-3px);
    box-shadow: 0 15px 40px rgba(34, 211, 238, 0.3);
}

/* Input */
textarea {
    background: rgba(0,0,0,0.35) !important;
    border: 1px solid var(--border) !important;
    border-radius: 16px !important;
    color: var(--text) !important;
    font-size: 16.5px !important;
    padding: 18px !important;
    min-height: 160px !important;
}
textarea:focus {
    border-color: var(--secondary) !important;
}

/* Output Labels (Cyan) */
label {
    color: var(--secondary) !important;
    font-weight: 600 !important;
    font-size: 15px !important;
}

/* Output Boxes - Subtle Dark */
.output-text {
    background: rgba(0,0,0,0.35) !important;
    border: 1px solid var(--border) !important;
    border-radius: 16px !important;
    padding: 20px !important;
    font-size: 18px !important;
    color: white !important;
    font-weight: 600;
    text-align: center;
}
"""

examples = [
    ["یہ بہت اچھا پروڈکٹ ہے"],
    ["مجھے یہ بالکل پسند نہیں آیا"],
    ["زبردست تجربہ تھا"],
    ["خدمات بہت خراب تھیں"],
    ["یہ ٹھیک تھا"]
]

with gr.Blocks(css=custom_css, title="Urdu Sentiment Analysis") as interface:
    
    gr.Markdown("# 🇵🇰 Urdu Sentiment <span>Analyzer</span>")
    gr.Markdown('<p class="subtitle">Fine-tuned mBERT for Accurate Urdu Sentiment Classification</p>')

    with gr.Row():
        with gr.Column(scale=7):
            text_input = gr.Textbox(
                label="Enter Urdu Text",
                placeholder="یہاں اردو جملہ لکھیں...",
                lines=9
            )
            
            analyze_btn = gr.Button("Analyze Sentiment")
            
            gr.Examples(
                examples=examples,
                inputs=text_input,
                label="Example Sentences"
            )

        with gr.Column(scale=5):
            gr.Markdown("### Prediction Results")
            
            sentiment = gr.Textbox(
                label="Predicted Sentiment",
                interactive=False,
                elem_classes="output-text"
            )
            
            confidence = gr.Textbox(
                label="Confidence Score",
                interactive=False,
                elem_classes="output-text"
            )

    # Footer
    gr.Markdown("""
    <div style="text-align:center; margin-top:50px; color:#6B7280; font-size:14px;">
        Built with Transformers + Gradio • Production-ready Urdu NLP Demo
    </div>
    """)

    analyze_btn.click(
        fn=predict_sentiment,
        inputs=text_input,
        outputs=[sentiment, confidence]
    )

interface.launch()



# from transformers import AutoTokenizer, AutoModelForSequenceClassification
# import gradio as gr
# import torch

# model_path = "./best_model"

# tokenizer = AutoTokenizer.from_pretrained(model_path)
# model = AutoModelForSequenceClassification.from_pretrained(model_path)

# device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# model.to(device)
# model.eval()


# def predict_sentiment(text):
#     inputs = tokenizer(
#         text,
#         return_tensors="pt",
#         truncation=True,
#         padding=True
#     ).to(device)

#     with torch.no_grad():
#         outputs = model(**inputs)

#     probs = torch.softmax(outputs.logits, dim=1)

#     pred_id = torch.argmax(probs, dim=1).item()
#     confidence = torch.max(probs).item()

#     label = model.config.id2label[pred_id]

#     return label, confidence




# # Gradio UI
# interface = gr.Interface(
#     fn=predict_sentiment,
#     inputs=gr.Textbox(label="Enter Urdu Text"),
#     outputs=[
#         gr.Textbox(label="Sentiment"),
#         gr.Textbox(label="Confidence")
#     ],
#     title="Urdu Sentiment Analysis (mBERT)",
# )

# interface.launch()