# -----------------------------------------------------------------------------
# Home-Ka — US real-estate aggregator (Groupe KA)
# Author: Simon-Pierre Boucher — contact@spboucher.ai
# quality.py : data-quality gate — completeness score and publication
# threshold (quarantine), same principle as immo-ka/quality.py.
# A listing without a plausible price, location, type or usable content stays
# out of the grid/map/stats until later cycles complete it; its detail page
# by uid remains reachable.
# -----------------------------------------------------------------------------
from __future__ import annotations

import sqlite3

PRICE_MIN, PRICE_MAX = 5_000, 200_000_000
PUBLISH_THRESHOLD = 0.55

_WEIGHTS = {
    "price": 0.30,       # plausible list price
    "location": 0.25,    # city+state or zip
    "address": 0.15,
    "type": 0.10,
    "content": 0.10,     # description or features
    "images": 0.10,
}


def score_row(r) -> float:
    s = 0.0
    price = r["list_price"]
    if price is not None and PRICE_MIN <= price <= PRICE_MAX:
        s += _WEIGHTS["price"]
    if (r["city"] and r["state"]) or r["zip_code"]:
        s += _WEIGHTS["location"]
    if r["street_address"]:
        s += _WEIGHTS["address"]
    if r["property_type"]:
        s += _WEIGHTS["type"]
    if (r["description"] and len(r["description"]) > 40) or \
            (r["features"] and r["features"] != "[]"):
        s += _WEIGHTS["content"]
    if r["images"] and r["images"] != "[]":
        s += _WEIGHTS["images"]
    return round(s, 3)


def refresh(con: sqlite3.Connection) -> dict:
    published = quarantined = 0
    for r in con.execute(
            "SELECT uid, list_price, city, state, zip_code, street_address,"
            " property_type, description, features, images"
            " FROM listings WHERE active=1"):
        q = score_row(r)
        # a plausible asking price is a HARD requirement for publication —
        # crawled sold/off-market shells must never reach the grid
        price_ok = (r["list_price"] is not None
                    and PRICE_MIN <= r["list_price"] <= PRICE_MAX)
        pub = 1 if (q >= PUBLISH_THRESHOLD and price_ok) else 0
        published += pub
        quarantined += 1 - pub
        con.execute("UPDATE listings SET quality=?, published=? WHERE uid=?",
                    (q, pub, r["uid"]))
    con.commit()
    return {"published": published, "quarantined": quarantined}


def summary(con: sqlite3.Connection) -> dict:
    rows = [dict(r) for r in con.execute(
        "SELECT source, COUNT(*) n, SUM(published) published,"
        " ROUND(AVG(quality), 3) avg_quality,"
        " SUM(CASE WHEN list_price IS NULL THEN 1 ELSE 0 END) no_price,"
        " SUM(CASE WHEN lat IS NULL THEN 1 ELSE 0 END) no_geo"
        " FROM listings WHERE active=1 GROUP BY source ORDER BY n DESC")]
    return {"by_source": rows}