app.py

from fastapi import FastAPI, File, UploadFile, Request
from fastapi.responses import HTMLResponse
from fastapi.staticfiles import StaticFiles
from fastapi.templating import Jinja2Templates

import os
import json
import base64
import re
from dotenv import load_dotenv
import requests
import torch
from diffusers import StableDiffusionPipeline
from PIL import Image
from io import BytesIO
from imghdr import what

load_dotenv()
api_key = os.getenv("GOOGLE_API")

app = FastAPI()

templates = Jinja2Templates(directory="templates")

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

os.environ["HF_HOME"] = "/tmp/huggingface"
os.environ["TRANSFORMERS_CACHE"] = "/tmp/huggingface/transformers"
os.environ["HF_DATASETS_CACHE"] = "/tmp/huggingface/datasets"
os.environ["DIFFUSERS_CACHE"] = "/tmp/huggingface/diffusers"

pipe = StableDiffusionPipeline.from_pretrained(
    "runwayml/stable-diffusion-v1-5",
    cache_dir="/tmp/huggingface",
    torch_dtype=torch.float16 if device == "cuda" else torch.float32
)

pipe.to(device)


def clean_filename(text):
    return re.sub(r'[^\w\-_\. ]', '_', text.strip().lower().replace(" ", "_"))


def generate_image_base64(food_name):
    prompt = f"Professional food photography of {food_name}, top-down view, realistic lighting"
    image = pipe(prompt).images[0]

    buffered = BytesIO()
    image.save(buffered, format="PNG")
    encoded_image = base64.b64encode(buffered.getvalue()).decode("utf-8")
    return encoded_image

def get_mime_type(image_bytes):
    kind = what(None, h=image_bytes)
    return f"image/{kind or 'jpeg'}"

def extract_menu_from_image(image_bytes):
    base64_image = base64.b64encode(image_bytes).decode('utf-8')
    url = f"https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-flash:generateContent?key={api_key}"
    
    prompt = """
    Extract the menu items from this image and return ONLY a JSON array like:
    [
      {
        "food": "Dish Name",
        "description": "Short description or empty string",
        "price": 10,
        "category": "Category"
      }
    ]
    """

    payload = {
        "contents": [
            {
                "parts": [
                    {"text": prompt},
                    {
                        "inline_data": {
                            "mime_type": get_mime_type(image_bytes),
                            "data": base64_image
                        }
                    }
                ]
            }
        ],
        "generationConfig": {
            "responseMimeType": "application/json"
        }
    }

    headers = {'Content-Type': 'application/json'}

    try:
        res = requests.post(url, headers=headers, json=payload)
        res.raise_for_status()
        text = res.json()['candidates'][0]['content']['parts'][0]['text']
        return json.loads(text)
    except Exception as e:
        print("Error extracting menu:", e)
        return []


@app.get("/", response_class=HTMLResponse)
async def form(request: Request):
    return templates.TemplateResponse("index.html", {"request": request})


@app.post("/upload", response_class=HTMLResponse)
async def upload(request: Request, menu_image: UploadFile = File(...)):
    image_bytes = await menu_image.read()
    menu_items = extract_menu_from_image(image_bytes)

    for item in menu_items:
        item["img_base64"] = generate_image_base64(item["food"])

    html = "<h2>🍽️ AI Food Menu</h2><div style='display:flex; flex-wrap:wrap;'>"
    for item in menu_items:
        html += f"""
        <div style='border:1px solid #ccc; margin:10px; width:220px; text-align:center; padding:10px; border-radius:10px; box-shadow:2px 2px 5px #aaa;'>
            <img src='data:image/png;base64,{item["img_base64"]}' width='200'><br>
            <h3>{item['food']}</h3>
            <p><b>${item['price']}</b></p>
            <p>{item['description']}</p>
        </div>
        """
    html += "</div><br><a href='/'>Upload Another</a>"
    return HTMLResponse(content=html)