app.py

import os
import torch
from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
import story_generator  # ✅ Import Story Generator
from diffusers import DiffusionPipeline
import io
import base64
from PIL import Image
from huggingface_hub import login

app = FastAPI()

# ✅ Set Hugging Face cache directory to /tmp
os.environ["HF_HOME"] = "/tmp/huggingface"
os.environ["TRANSFORMERS_CACHE"] = "/tmp/huggingface"
os.environ["HF_HUB_CACHE"] = "/tmp/huggingface"

# ✅ Hugging Face Authentication (Only needed if model is private)
HF_TOKEN = "your_huggingface_token_here"  # Replace this with your actual token
login(token=HF_TOKEN)

# ✅ Load Image Generation Model (Use a fast, public model)
IMAGE_MODEL = "stabilityai/sdxl-turbo"  # Replace with "stabilityai/sdxl-lightning" if needed
pipeline = DiffusionPipeline.from_pretrained(
    IMAGE_MODEL,
    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
    use_auth_token=HF_TOKEN  # Required for private models
).to("cuda" if torch.cuda.is_available() else "cpu")

# ✅ Define the input request format
class StoryRequest(BaseModel):
    theme: str
    reading_level: str

@app.post("/generate_story_questions_images")
def generate_story_questions_images(request: StoryRequest):
    """Generates a story, questions, and corresponding images."""
    try:
        print(f"🎭 Generating story for theme: {request.theme} and level: {request.reading_level}")

        # ✅ Generate the story and questions
        if not hasattr(story_generator, "generate_story_and_questions"):
            raise HTTPException(status_code=500, detail="Story generation function not found in story_generator.py")

        story_result = story_generator.generate_story_and_questions(request.theme, request.reading_level)
        story_text = story_result["story"]
        questions = story_result["questions"]

        # ✅ Split the story into sentences for image generation
        story_sentences = story_text.strip().split(". ")

        # ✅ Generate an image for each sentence
        images = []
        for sentence in story_sentences:
            if len(sentence) > 5:  # Avoid empty sentences
                print(f"🖼️ Generating image for: {sentence}")
                image = pipeline(prompt=sentence, num_inference_steps=5).images[0]

                # Convert Image to Base64
                img_byte_arr = io.BytesIO()
                image.save(img_byte_arr, format="PNG")
                img_base64 = base64.b64encode(img_byte_arr.getvalue()).decode("utf-8")

                images.append({"sentence": sentence, "image": img_base64})

        # ✅ Return the full response
        return {
            "theme": request.theme,
            "reading_level": request.reading_level,
            "story": story_text,
            "questions": questions,
            "images": images,
        }

    except Exception as e:
        print(f"❌ Error generating story, questions, or images: {e}")
        raise HTTPException(status_code=500, detail=str(e))

# ✅ Welcome message at root
@app.get("/")
def home():
    return {"message": "🎉 Welcome to the Story, Question & Image Generation API! Use /generate_story_questions_images"}