studspider/studspider.py

import os
import json
import asyncio
from typing import Dict, Any
from fastapi import FastAPI, Request, Response, status
from fastapi.responses import JSONResponse
import src.get_sq_env as sq_env
import src.stats as stats

# Square webhook validation
from square.utils.webhooks_helper import verify_signature

# Configuration
SQUARE_WEBHOOK_SIGNATURE_KEY = sq_env.SQUARE_WEBHOOK_SIGNATURE_KEY
SQUARE_WEBHOOK_URL = sq_env.SQUARE_WEBHOOK_URL
SQUARE_LEGO_CAT_ID = sq_env.LEGO_CAT_ID

# In-memory store for processed event IDs (idempotency)
processed_event_ids = set()
MAX_STORED_EVENT_IDS = 10000

# Create FastAPI app
app = FastAPI(title="Studspider")


@app.get("/")
async def root():
    """Landing page"""
    return {
        "service": "Studspider",
        "status": "running",
        "endpoints": {
            "webhook": "/hook (POST)",
            "health": "/health (GET)",
            "stats": "/stats (GET)",
        },
    }


@app.get("/health")
async def health_check():
    """Health check endpoint for monitoring"""
    return {
        "status": "healthy",
        "signature_verification": (
            "enabled" if SQUARE_WEBHOOK_SIGNATURE_KEY else "disabled"
        ),
    }


@app.get("/stats")
async def get_stats():
    """
    Get daily statistics for LEGO inventory.

    Returns metrics about enrichment progress, missing data, catalog value, and cache stats.
    """
    try:
        daily_stats = await stats.get_daily_stats()
        cache_stats = await stats.get_brickset_cache_stats()

        # Calculate completion percentage
        if daily_stats["total_items"] > 0:
            completion_rate = (
                daily_stats["enriched_items"] / daily_stats["total_items"] * 100
            )
        else:
            completion_rate = 0.0

        return {
            "status": "success",
            "version": "0.3.0",
            "git_commit": stats.get_git_commit(),
            "stats": daily_stats,
            "completion_rate": round(completion_rate, 1),
            "brickset_cache": cache_stats,
        }
    except Exception as e:
        return JSONResponse(
            status_code=500,
            content={"status": "error", "message": str(e)},
        )


@app.post("/hook")
async def webhook_handler(request: Request):
    """
    Handle Square webhook events - MUST respond within 2 seconds!

    Square sends POST requests with:
    - x-square-hmacsha256-signature header
    - JSON body with event data
    - event_id for idempotency
    """

    try:
        # Log incoming request
        print(f"\n[WEBHOOK] Incoming request from {request.client.host}")
        print(f"[WEBHOOK] Method: {request.method}")
        print(f"[WEBHOOK] URL: {request.url}")
        print(f"[WEBHOOK] Headers: {dict(request.headers)}")

        # Read raw body for signature verification
        body = await request.body()
        body_str = body.decode("utf-8")
        print(f"[WEBHOOK] Body length: {len(body_str)} bytes")

        # Get signature header
        signature_header = request.headers.get("x-square-hmacsha256-signature", "")

        # Verify Square webhook signature (if configured)
        if SQUARE_WEBHOOK_SIGNATURE_KEY and SQUARE_WEBHOOK_URL:
            if not signature_header:
                print(
                    f"[ERROR] Missing x-square-hmacsha256-signature header from {request.client.host}"
                )
                return JSONResponse(
                    status_code=status.HTTP_401_UNAUTHORIZED,
                    content={"error": "Missing signature header"},
                )

            # Debug signature verification inputs
            print(f"[DEBUG] Signature header: {signature_header[:20]}...")
            print(
                f"[DEBUG] Signature key (first 10 chars): {SQUARE_WEBHOOK_SIGNATURE_KEY[:10]}..."
            )
            print(f"[DEBUG] Notification URL: {SQUARE_WEBHOOK_URL}")
            print(f"[DEBUG] Body (first 100 chars): {body_str[:100]}...")

            try:
                is_valid = verify_signature(
                    request_body=body_str,
                    signature_header=signature_header,
                    signature_key=SQUARE_WEBHOOK_SIGNATURE_KEY,
                    notification_url=SQUARE_WEBHOOK_URL,
                )

                if not is_valid:
                    print(
                        f"[ERROR] Invalid webhook signature from {request.client.host}"
                    )
                    print(f"[ERROR] This usually means:")
                    print(
                        f"[ERROR]   1. SQUARE_WEBHOOK_URL doesn't match Square Dashboard exactly"
                    )
                    print(f"[ERROR]   2. SQUARE_WEBHOOK_SIGNATURE_KEY is incorrect")
                    print(f"[ERROR]   3. Request body was modified in transit")
                    return JSONResponse(
                        status_code=status.HTTP_401_UNAUTHORIZED,
                        content={"error": "Invalid signature"},
                    )

                print(f"[INFO] ✓ Webhook signature verified from {request.client.host}")
            except Exception as e:
                print(f"[ERROR] Signature verification exception: {e}")
                return JSONResponse(
                    status_code=status.HTTP_401_UNAUTHORIZED,
                    content={"error": f"Signature verification failed: {str(e)}"},
                )
        else:
            print(
                f"[WARN] Webhook signature verification DISABLED - configure SQUARE_WEBHOOK_SIGNATURE_KEY"
            )

        # Parse JSON data
        try:
            data = json.loads(body_str)
        except json.JSONDecodeError as e:
            print(f"[ERROR] Invalid JSON from {request.client.host}: {e}")
            return JSONResponse(
                status_code=status.HTTP_400_BAD_REQUEST,
                content={"error": "Invalid JSON"},
            )

        # Check idempotency (event_id)
        event_id = data.get("event_id")
        if event_id:
            if event_id in processed_event_ids:
                print(
                    f"[INFO] Duplicate event {event_id} - already processed (idempotency)"
                )
                return JSONResponse(
                    status_code=status.HTTP_200_OK,
                    content={
                        "status": "success",
                        "message": "Event already processed",
                        "event_id": event_id,
                    },
                )

            # Add to processed set
            processed_event_ids.add(event_id)

            # Prevent memory growth
            if len(processed_event_ids) > MAX_STORED_EVENT_IDS:
                processed_event_ids.pop()

        # Get event info for quick response
        webhook_type = data.get("type", "unknown")
        event_id_val = data.get("event_id", "none")
        merchant_id = data.get("merchant_id", "unknown")

        # RESPOND IMMEDIATELY - Square requires fast response (< 2 seconds)
        response_data = {
            "status": "success",
            "message": "Webhook received",
            "event_id": event_id_val,
        }

        # Process webhook in background (async task)
        asyncio.create_task(
            process_webhook(
                webhook_type, event_id_val, merchant_id, data, request.client.host
            )
        )

        # Return immediate response
        print(f"[INFO] 200 OK response sent for event {event_id_val}")
        return JSONResponse(status_code=status.HTTP_200_OK, content=response_data)

    except Exception as e:
        # Catch-all for any unexpected errors - still respond quickly
        print(f"[ERROR] Unexpected error in webhook handler: {e}")
        import traceback

        traceback.print_exc()
        return JSONResponse(
            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
            content={"error": "Internal server error"},
        )

    # Return immediate response
    return JSONResponse(status_code=status.HTTP_200_OK, content=response_data)


async def process_webhook(
    webhook_type: str,
    event_id: str,
    merchant_id: str,
    data: Dict[str, Any],
    client_ip: str,
):
    """
    Process webhook data asynchronously after responding to Square

    This runs in the background and doesn't block the HTTP response
    """
    try:
        print("\n" + "=" * 60)
        print(f"[WEBHOOK] Received at {asyncio.get_event_loop().time()}")
        print(f"[WEBHOOK] From: {client_ip}")
        print(f"[WEBHOOK] Type: {webhook_type}")
        print(f"[WEBHOOK] Event ID: {event_id}")
        print(f"[WEBHOOK] Merchant: {merchant_id}")
        print(f"[WEBHOOK] Full payload:")
        print(json.dumps(data, indent=2))
        print("=" * 60 + "\n")

        # Process based on webhook type
        if webhook_type == "inventory.count.updated":
            print(f"[PROCESS] Processing inventory count update...")
            inventory_counts = (
                data.get("data", {}).get("object", {}).get("inventory_counts", [])
            )

            # Import here to avoid circular imports
            import src.asquare as asquare
            import src.glue as glue

            for count in inventory_counts:
                variation_id = count.get("catalog_object_id")
                quantity = count.get("quantity")
                state = count.get("state")
                print(
                    f"[PROCESS]   Catalog: {variation_id}, Qty: {quantity}, State: {state}"
                )

                try:
                    # Get parent item from variation ID
                    item = await asquare.get_item_by_variation_id(variation_id)
                    item_id = item.get("id")
                    item_name = item.get("name", "")
                    category_id = item.get("category_id")

                    print(f"[PROCESS]   Item: {item_id} - {item_name}")
                    print(f"[PROCESS]   Category: {category_id}")

                    # Only process LEGO category items
                    if category_id == SQUARE_LEGO_CAT_ID:
                        print(
                            f"[PROCESS]   ✓ LEGO item detected - triggering enrichment"
                        )
                        # Trigger enrichment in background with item data to avoid redundant lookup
                        asyncio.create_task(
                            glue.enrich_single_item(item_id, item_data=item)
                        )
                    else:
                        print(f"[PROCESS]   ✗ Not a LEGO item - skipping enrichment")

                except Exception as e:
                    print(f"[ERROR] Failed to process variation {variation_id}: {e}")

        elif webhook_type == "catalog.version.updated":
            print(f"[PROCESS] Processing catalog version update...")

            # Import here to avoid circular imports
            import src.asquare as asquare
            import src.glue as glue

            catalog_object = data.get("data", {}).get("object", {})
            object_type = catalog_object.get("type")

            if object_type == "ITEM":
                item_id = catalog_object.get("id")
                print(f"[PROCESS]   Item updated: {item_id}")

                try:
                    # Get item details to check category
                    item = await asquare.get_item_by_id(item_id)
                    item_name = item.get("name", "")
                    category_id = item.get("category_id")

                    print(f"[PROCESS]   Item: {item_id} - {item_name}")
                    print(f"[PROCESS]   Category: {category_id}")

                    # Only process LEGO category items
                    if category_id == SQUARE_LEGO_CAT_ID:
                        print(
                            f"[PROCESS]   ✓ LEGO item detected - triggering enrichment"
                        )
                        # Trigger enrichment in background with item data to avoid redundant lookup
                        asyncio.create_task(
                            glue.enrich_single_item(item_id, item_data=item)
                        )
                    else:
                        print(f"[PROCESS]   ✗ Not a LEGO item - skipping enrichment")

                except Exception as e:
                    print(f"[ERROR] Failed to process item {item_id}: {e}")
            else:
                print(f"[PROCESS]   Object type '{object_type}' - skipping")

        else:
            print(f"[PROCESS] Webhook type '{webhook_type}' - no handler configured")

        print(f"[PROCESS] ✓ Webhook processing complete for {event_id}")

    except Exception as e:
        print(f"[ERROR] Failed to process webhook {event_id}: {e}")
        import traceback

        traceback.print_exc()


if __name__ == "__main__":
    # For development/testing - use uvicorn directly
    import uvicorn

    print("\n" + "=" * 60)
    print("Studspider webhook boot dev")
    print("=" * 60)
    print(
        f"Signature verification: {'ENABLED' if SQUARE_WEBHOOK_SIGNATURE_KEY else 'DISABLED'}"
    )
    print(f"Webhook URL: {SQUARE_WEBHOOK_URL or 'NOT CONFIGURED'}")
    print("\nFor production, use Gunicorn:")
    print(
        "  gunicorn conf/studspider:app -w 4 -k uvicorn.workers.UvicornWorker --bind 0.0.0.0:7433"
    )
    print("=" * 60 + "\n")

    uvicorn.run(app, host="0.0.0.0", port=7433, log_level="info")