from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer

# Define the input schema
class ModelInput(BaseModel):
    prompt: str
    max_new_tokens: int = 50  # Optional: Defaults to 50 tokens

# Initialize FastAPI app
app = FastAPI()

# Load your model and tokenizer
model_path = "khurrameycon/SmolLM-135M-Instruct-qa_pairs_converted.json-25epochs"
tokenizer = AutoTokenizer.from_pretrained(model_path)
model = AutoModelForCausalLM.from_pretrained(model_path)

# Initialize the pipeline
generator = pipeline("text-generation", model=model, tokenizer=tokenizer)

# Helper function to generate a response
def generate_response(model, tokenizer, instruction, max_new_tokens=128):
    """Generate a response from the model based on an instruction."""
    try:
        # Format the input as chat messages if necessary
        messages = [{"role": "user", "content": instruction}]
        input_text = tokenizer.apply_chat_template(
            messages, tokenize=False, add_generation_prompt=True
        )
        # Tokenize and generate the output
        inputs = tokenizer.encode(input_text, return_tensors="pt")
        outputs = model.generate(
            inputs,
            max_new_tokens=max_new_tokens,
            temperature=0.2,
            top_p=0.9,
            do_sample=True,
        )
        # Decode the output
        response = tokenizer.decode(outputs[0], skip_special_tokens=True)
        return response
    except Exception as e:
        raise ValueError(f"Error generating response: {e}")

@app.post("/generate")
def generate_text(input: ModelInput):
    """API endpoint to generate text."""
    try:
        # Call the helper function
        response = generate_response(
            model=model, tokenizer=tokenizer, instruction=input.prompt, max_new_tokens=input.max_new_tokens
        )
        return {"generated_text": response}
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

@app.get("/")
def root():
    return {"message": "Welcome to the Hugging Face Model API!"}