"""
Read-only audit for listings missing usable geolocation.

Prints listings where latitude/longitude are missing, null, blank,
non-numeric, or outside valid coordinate ranges.
"""

import asyncio
from typing import Any, Optional

from app.database import connect_db, disconnect_db, get_db


def _to_float(value: Any) -> Optional[float]:
    if value is None:
        return None
    if isinstance(value, str) and not value.strip():
        return None
    try:
        return float(value)
    except (TypeError, ValueError):
        return None


def _geo_problem(doc: dict) -> Optional[str]:
    has_lat = "latitude" in doc
    has_lon = "longitude" in doc
    lat = _to_float(doc.get("latitude"))
    lon = _to_float(doc.get("longitude"))

    if not has_lat and not has_lon:
        return "missing latitude and longitude fields"
    if not has_lat:
        return "missing latitude field"
    if not has_lon:
        return "missing longitude field"
    if lat is None and lon is None:
        return "empty/invalid latitude and longitude"
    if lat is None:
        return "empty/invalid latitude"
    if lon is None:
        return "empty/invalid longitude"
    if not -90 <= lat <= 90:
        return "latitude out of range"
    if not -180 <= lon <= 180:
        return "longitude out of range"
    return None


async def main():
    await connect_db()
    try:
        db = await get_db()
        total = await db.listings.count_documents({})
        active_total = await db.listings.count_documents({"status": "active"})

        projection = {
            "title": 1,
            "location": 1,
            "address": 1,
            "status": 1,
            "listing_type": 1,
            "latitude": 1,
            "longitude": 1,
            "created_at": 1,
            "createdAt": 1,
        }
        cursor = db.listings.find({}, projection).sort("created_at", -1)

        missing = []
        async for doc in cursor:
            problem = _geo_problem(doc)
            if problem:
                missing.append((problem, doc))

        print("=== Listings Missing Usable Geolocation ===")
        print(f"Total listings: {total}")
        print(f"Active listings: {active_total}")
        print(f"Need geolocation: {len(missing)}")
        print()

        by_status = {}
        by_problem = {}
        for problem, doc in missing:
            by_status[doc.get("status") or "unknown"] = by_status.get(doc.get("status") or "unknown", 0) + 1
            by_problem[problem] = by_problem.get(problem, 0) + 1

        print("By status:")
        for status, count in sorted(by_status.items()):
            print(f"  {status}: {count}")
        print()

        print("By problem:")
        for problem, count in sorted(by_problem.items()):
            print(f"  {problem}: {count}")
        print()

        for idx, (problem, doc) in enumerate(missing, start=1):
            title = (doc.get("title") or "No title").replace("\n", " ")[:80]
            print(f"{idx}. {title}")
            print(f"   id: {doc.get('_id')}")
            print(f"   status: {doc.get('status') or 'unknown'}")
            print(f"   type: {doc.get('listing_type') or 'unknown'}")
            print(f"   location: {doc.get('location')}")
            print(f"   address: {doc.get('address')}")
            print(f"   latitude: {doc.get('latitude')!r}")
            print(f"   longitude: {doc.get('longitude')!r}")
            print(f"   problem: {problem}")
            print()
    finally:
        await disconnect_db()


if __name__ == "__main__":
    asyncio.run(main())