gendesign/tradein-mvp/backend/app/services/matching/houses.py

"""House cross-source matching — 3-tier algorithm.

Tier 0 (confidence 1.0): cadastral_number exact match on houses table.
Tier 1 (confidence 1.0): ext_source + ext_id already in house_sources.
Tier 2 (confidence 0.9): address_fingerprint match in house_address_aliases.
Tier 3 (confidence 0.7): geo-proximity within 30 m (PostGIS ST_DWithin).
New   (confidence 1.0): INSERT new canonical house.

Algorithm reference: decisions/Cross_Source_Matching_Strategy.md sec 3
"""

import logging

from sqlalchemy import text
from sqlalchemy.orm import Session

from app.services.matching.normalize import address_fingerprint, normalize_address

logger = logging.getLogger(__name__)


def match_or_create_house(
    db: Session,
    ext_source: str,
    ext_id: str,
    address: str | None = None,
    lat: float | None = None,
    lon: float | None = None,
    *,
    year_built: int | None = None,
    building_cadastral_number: str | None = None,
    cadastral_number: str | None = None,
    source_url: str | None = None,
) -> tuple[int, float, str]:
    """Match existing house or create new canonical record.

    ⚠️ KNOWN LIMITATION: Concurrent calls for the same address can race —
    two scrapers calling simultaneously for an unknown address can both
    miss Tier 0-3 and each INSERT a new house row. Mitigation (Stage 8.x):
    gate scraper concurrency to 1 per fingerprint OR add
    pg_advisory_xact_lock(hashtext(normalized_address)) at function entry.

    Returns:
        (house_id, confidence ∈ [0.0, 1.0], method ∈ {
            'cadastr_exact', 'source_exact', 'fingerprint', 'geo_proximity', 'new'
        })

    Method values:
        'cadastr_exact'  — matched by cadastral number (confidence 1.0)
        'source_exact'   — already in house_sources for this source+ext_id (confidence 1.0)
        'fingerprint'    — matched by address fingerprint (confidence 0.9)
        'geo_proximity'  — matched by geo within 30 m (confidence 0.7)
        'new'            — new house created (confidence 1.0)
    """
    cad = building_cadastral_number or cadastral_number

    # Tier 0: cadastral number exact match
    if cad:
        row = db.execute(
            text('SELECT id FROM houses WHERE cadastral_number = :cad ORDER BY id ASC LIMIT 1'),
            {'cad': cad},
        ).mappings().first()
        if row:
            house_id = int(row['id'])
            _upsert_house_source(
                db, house_id=house_id, ext_source=ext_source, ext_id=ext_id,
                method='cadastr_exact', confidence=1.0,
            )
            logger.info('house match cadastr_exact house_id=%s cad=%s', house_id, cad)
            return (house_id, 1.0, 'cadastr_exact')

    # Tier 1: source+ext_id already registered in house_sources
    row = db.execute(
        text(
            'SELECT house_id FROM house_sources '
            'WHERE ext_source = :s AND ext_id = :e LIMIT 1'
        ),
        {'s': ext_source, 'e': str(ext_id)},
    ).mappings().first()
    if row:
        house_id = int(row['house_id'])
        logger.info('house match source_exact house_id=%s src=%s ext_id=%s', house_id, ext_source,
                    ext_id)
        return (house_id, 1.0, 'source_exact')

    # Compute fingerprint once for Tier 2 and reuse for new house registration
    fp = address_fingerprint(address, lat, lon)

    # Tier 2: address fingerprint lookup
    row = db.execute(
        text(
            'SELECT house_id FROM house_address_aliases '
            'WHERE fingerprint = :fp LIMIT 1'
        ),
        {'fp': fp},
    ).mappings().first()
    if row:
        house_id = int(row['house_id'])
        _upsert_house_source(
            db, house_id=house_id, ext_source=ext_source, ext_id=ext_id,
            method='fingerprint', confidence=0.9,
        )
        logger.info('house match fingerprint house_id=%s fp=%s', house_id, fp)
        return (house_id, 0.9, 'fingerprint')

    # Tier 3: geo-proximity — within 30 m (PostGIS geography cast on both sides)
    if lat is not None and lon is not None:
        row = db.execute(
            text("""
                SELECT id,
                       ST_Distance(geom::geography, ST_MakePoint(:lon, :lat)::geography) AS dist
                FROM houses
                WHERE geom IS NOT NULL
                  AND ST_DWithin(geom::geography, ST_MakePoint(:lon, :lat)::geography, 30)
                ORDER BY dist ASC
                LIMIT 1
            """),
            {'lat': lat, 'lon': lon},
        ).mappings().first()
        if row:
            house_id = int(row['id'])
            _upsert_house_source(
                db, house_id=house_id, ext_source=ext_source, ext_id=ext_id,
                method='geo_proximity', confidence=0.7,
            )
            # Register fingerprint alias so future calls skip geo lookup
            _insert_alias(db, house_id=house_id, address=address, fp=fp, source=ext_source)
            logger.info(
                'house match geo_proximity house_id=%s dist=%.1f src=%s',
                house_id, float(row['dist']), ext_source,
            )
            return (house_id, 0.7, 'geo_proximity')

    # New house — INSERT canonical record.
    # geom column is auto-populated by houses_set_geom_trg BEFORE INSERT trigger from lat/lon.
    # Do NOT include geom in the INSERT column list — trigger handles it.
    url = source_url or f'matching://{ext_source}/{ext_id}'
    row = db.execute(
        text("""
            INSERT INTO houses (source, ext_house_id, url, address, lat, lon, year_built,
                                cadastral_number)
            VALUES (
                :src, :eid, :url,
                :addr,
                CAST(:lat AS double precision),
                CAST(:lon AS double precision),
                CAST(:yb AS integer),
                :cad
            )
            ON CONFLICT (source, ext_house_id) DO UPDATE SET
                address = COALESCE(EXCLUDED.address, houses.address)
            RETURNING id
        """),
        {
            'src': ext_source,
            'eid': str(ext_id),
            'url': url,
            'addr': address,
            'lat': lat,
            'lon': lon,
            'yb': year_built,
            'cad': cad,
        },
    ).mappings().one()
    house_id = int(row['id'])

    _upsert_house_source(
        db, house_id=house_id, ext_source=ext_source, ext_id=ext_id,
        method='new', confidence=1.0,
    )
    _insert_alias(db, house_id=house_id, address=address, fp=fp, source=ext_source)

    logger.info('house new house_id=%s addr=%r src=%s', house_id, address, ext_source)
    return (house_id, 1.0, 'new')


def _upsert_house_source(
    db: Session,
    *,
    house_id: int,
    ext_source: str,
    ext_id: str,
    method: str,
    confidence: float,
) -> None:
    """Insert or refresh house_sources row for this source+ext_id."""
    db.execute(
        text("""
            INSERT INTO house_sources (
                house_id, ext_source, ext_id, confidence, matched_method, matched_at, last_seen_at
            ) VALUES (
                CAST(:hid AS bigint), :s, :e,
                CAST(:c AS real), :m, NOW(), NOW()
            )
            ON CONFLICT (ext_source, ext_id) DO UPDATE SET
                confidence    = GREATEST(EXCLUDED.confidence, house_sources.confidence),
                last_seen_at  = NOW()
        """),
        {'hid': house_id, 's': ext_source, 'e': str(ext_id), 'c': confidence, 'm': method},
    )


def _insert_alias(
    db: Session,
    *,
    house_id: int,
    address: str | None,
    fp: str,
    source: str,
) -> None:
    """Register normalized_address alias for this house (idempotent on normalized_address)."""
    db.execute(
        text("""
            INSERT INTO house_address_aliases (house_id, normalized_address, fingerprint, source)
            VALUES (CAST(:hid AS bigint), :na, :fp, :src)
            ON CONFLICT (normalized_address) DO NOTHING
        """),
        {
            'hid': house_id,
            'na': normalize_address(address),
            'fp': fp,
            'src': source,
        },
    )