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

MODEL_NAME = "Devishetty100/redmoon-gibberishdetective"

try:
    # Try loading with pipeline first
    classifier = pipeline(
        "text-classification",
        model=MODEL_NAME
    )
except:
    try:
        # Fallback: Load components separately
        tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased")
        model = AutoModelForSequenceClassification.from_pretrained(MODEL_NAME)
        classifier = pipeline(
            "text-classification",
            model=model,
            tokenizer=tokenizer,
            device=0 if torch.cuda.is_available() else -1
        )
    except Exception as e:
        raise gr.Error(f"Failed to load model: {str(e)}\n\nPlease ensure all model files exist in the repository.")

def predict(text):
    try:
        if not text.strip():
            return {"Error": "Please enter some text"}
        
        result = classifier(text)[0]
        return {
            "Prediction": result['label'],
            "Confidence": f"{result['score']:.2%}"
        }
    except Exception as e:
        return {"Error": str(e)}

with gr.Blocks(theme=gr.themes.Soft()) as demo:
    gr.Markdown("""
    # 🔍 Gibberish Detective
    *Using model: {MODEL_NAME}*
    """)
    
    with gr.Row():
        input_text = gr.Textbox(label="Input Text", placeholder="Enter text here...", lines=3)
        output_json = gr.JSON(label="Results")
    
    gr.Examples(
        examples=[
            ["This is proper English"],
            ["Xkjsd hfkjshdf 9834 kjsdhf!"],
            ["Hello world"],
            ["Hfjsd kjsadf lkjsdf 1234!"]
        ],
        inputs=input_text
    )
    
    input_text.submit(predict, inputs=input_text, outputs=output_json)
    gr.Button("Analyze").click(predict, inputs=input_text, outputs=output_json)

if __name__ == "__main__":
    demo.launch()