app.py

#!/usr/bin/env python3
"""
NeuralMesh Backend - CID Metadata + Hugging Face Execution
Sepolia + Pinata + Hugging Face
"""

import os
import time
import logging
from contextlib import asynccontextmanager

from dotenv import load_dotenv
from fastapi import FastAPI, HTTPException
from fastapi.middleware.cors import CORSMiddleware
from pydantic import BaseModel
from web3 import Web3
from eth_account import Account
import httpx
from sentence_transformers import SentenceTransformer
# ==========================================
# LOAD ENV
# ==========================================
load_dotenv()

CONFIG = {
    "RPC_URL": "https://rpc.sepolia.org",
    "CONTRACT_ADDRESS": "0xc76Bf13d48C61A68865aa16D91D2ECf86e7Fc773",
    "PRIVATE_KEY": os.getenv("ORACLE_PRIVATE_KEY"),
    "CHAIN_ID": 11155111,
    "PINATA_GATEWAY": "https://gateway.pinata.cloud/ipfs/",
    "HF_TOKEN": os.getenv("HF_TOKEN"),
    "HOST": "0.0.0.0",
    "PORT": 8000,
    "CORS_ORIGINS": ["http://localhost:5173"],
}

if not CONFIG["PRIVATE_KEY"]:
    raise ValueError("❌ ORACLE_PRIVATE_KEY not set")

if not CONFIG["HF_TOKEN"]:
    raise ValueError("❌ HF_TOKEN not set")

logging.basicConfig(
    level=logging.INFO,
    format="%(asctime)s | %(levelname)-8s | %(message)s"
)
logger = logging.getLogger("NeuralMesh")

# ==========================================
# CONTRACT ABI
# ==========================================
CONTRACT_ABI = [
    {
        "inputs": [{"name": "_modelId", "type": "uint256"}, {"name": "_user", "type": "address"}],
        "name": "hasAccess",
        "outputs": [{"name": "", "type": "bool"}],
        "stateMutability": "view",
        "type": "function"
    },
    {
        "inputs": [{"name": "_modelId", "type": "uint256"}],
        "name": "models",
        "outputs": [
            {"name": "model_id", "type": "uint256"},
            {"name": "model_cid", "type": "string"},
            {"name": "model_creator", "type": "address"},
            {"name": "model_price", "type": "uint256"},
            {"name": "model_usageCount", "type": "uint256"},
            {"name": "is_model_exists", "type": "bool"}
        ],
        "stateMutability": "view",
        "type": "function"
    }
]

# ==========================================
# BLOCKCHAIN SERVICE
# ==========================================
class BlockchainService:
    def __init__(self):
        self.w3 = Web3(Web3.HTTPProvider(CONFIG["RPC_URL"]))

        if not self.w3.is_connected():
            raise ConnectionError("❌ Cannot connect to Sepolia RPC")

        self.contract = self.w3.eth.contract(
            address=Web3.to_checksum_address(CONFIG["CONTRACT_ADDRESS"]),
            abi=CONTRACT_ABI
        )

        self.account = Account.from_key(CONFIG["PRIVATE_KEY"])

        logger.info(f"🔗 Connected to Sepolia")
        logger.info(f"🔐 Oracle: {self.account.address}")

    def verify_access(self, model_id: int, user_address: str) -> bool:
        try:
            user = Web3.to_checksum_address(user_address)
            has_access = self.contract.functions.hasAccess(model_id, user).call()
            model = self.contract.functions.models(model_id).call()
            return has_access and model[5]
        except Exception as e:
            logger.error(f"Access verification failed: {e}")
            return False

    def get_model_cid(self, model_id: int):
        try:
            model = self.contract.functions.models(model_id).call()
            if model[5]:
                return model[1]
            return None
        except Exception as e:
            logger.error(f"CID fetch failed: {e}")
            return None

# ==========================================
# MODEL SERVICE (HUGGING FACE EXECUTION)
# ==========================================

class ModelService:
    def __init__(self):
        self.model = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2")

    async def run_inference(self, model_repo: str, text: str):
        return self.model.encode(text).tolist()
# ==========================================
# METADATA FETCH (IPFS)
# ==========================================
async def fetch_metadata(cid: str):
    url = f"{CONFIG['PINATA_GATEWAY']}{cid}"
    async with httpx.AsyncClient(timeout=30.0) as client:
        response = await client.get(url)
        response.raise_for_status()
        return response.json()

# ==========================================
# FASTAPI
# ==========================================
class InferenceRequest(BaseModel):
    model_id: int
    user_address: str
    text: str

blockchain_service = None
model_service = ModelService()

@asynccontextmanager
async def lifespan(app: FastAPI):
    global blockchain_service
    logger.info("🚀 Starting NeuralMesh Backend...")
    blockchain_service = BlockchainService()
    yield

app = FastAPI(title="NeuralMesh", version="2.0.0", lifespan=lifespan)

app.add_middleware(
    CORSMiddleware,
    allow_origins=CONFIG["CORS_ORIGINS"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

# ==========================================
# ROUTES
# ==========================================
@app.get("/")
def root():
    return {
        "service": "NeuralMesh",
        "network": "Sepolia",
        "contract": CONFIG["CONTRACT_ADDRESS"],
        "oracle": blockchain_service.account.address if blockchain_service else None
    }

@app.get("/health")
def health():
    return {
        "status": "healthy",
        "blockchain_connected": blockchain_service.w3.is_connected(),
        "oracle": blockchain_service.account.address,
        "hf_token_configured": bool(CONFIG["HF_TOKEN"])
    }

@app.post("/embed")
async def embed(req: InferenceRequest):
    start = time.time()

    # 1️⃣ Verify ownership
    if not blockchain_service.verify_access(req.model_id, req.user_address):
        raise HTTPException(status_code=403, detail="Purchase required")
    print("come here ")

    # 2️⃣ Get CID from contract
    cid = blockchain_service.get_model_cid(req.model_id)
    print("the cid is ", cid)
    if not cid:
        raise HTTPException(status_code=404, detail="Model not found")

    # 3️⃣ Fetch metadata from IPFS
    metadata = await fetch_metadata(cid)
    print("metadata is ",metadata)
    # hf_repo = metadata.get("huggingFaceUrl")
    hf_repo = "sentence-transformers/all-MiniLM-L6-v2"
    print("hf repo is ",hf_repo)


    if not hf_repo:
        raise HTTPException(status_code=500, detail="Invalid metadata")

    # 4️⃣ Call Hugging Face
    output = await model_service.run_inference(hf_repo, req.text)
    print("the output is ",output)
    return {
        "model_id": req.model_id,
        "output": output,
        "processing_time_ms": round((time.time() - start) * 1000, 2)
    }

# ==========================================
# RUN SERVER
# ==========================================
if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host=CONFIG["HOST"], port=CONFIG["PORT"])