#!/usr/bin/env python3
"""Sync only explicitly mapped OpenStreetMap entrances into place_gates.

Entrances are extracted from the same OSM PBF used for the routing/map update.
The importer associates a node only when it is covered by exactly one named
target facility polygon (hospital/clinic, mall, or government facility).
No coordinates or gate types are inferred from facility centroids.
"""

from __future__ import annotations

import argparse
import csv
import json
import os
import re
import subprocess
import sys
import tempfile
from collections import Counter
from pathlib import Path


APP_DIR = Path(__file__).resolve().parents[1]
COUNTRY_NAMES = {
    "AE": "الإمارات", "BH": "البحرين", "EG": "مصر", "IQ": "العراق",
    "JO": "الأردن", "KW": "الكويت", "OM": "عُمان", "QA": "قطر",
    "SA": "السعودية", "SY": "سوريا", "YE": "اليمن",
}
USEFUL_ENTRANCES = {
    "yes", "main", "secondary", "service", "emergency", "parking", "entrance"
}


def run(command: list[str], *, input_text: str | None = None) -> str:
    proc = subprocess.run(
        command, cwd=APP_DIR, input=input_text, text=True,
        stdout=subprocess.PIPE, stderr=subprocess.PIPE,
    )
    if proc.returncode:
        sys.stderr.write(proc.stderr)
        raise RuntimeError(f"Command failed ({proc.returncode}): {' '.join(command)}")
    return proc.stdout


def psql(sql: str) -> str:
    return run([
        "docker", "compose", "exec", "-T", "db", "psql", "-X", "-v",
        "ON_ERROR_STOP=1", "-U", os.getenv("DB_USER", "mapuser"), "-d",
        os.getenv("DB_NAME", "mapdb"), "-At", "-c", sql,
    ])


def label_for(tags: dict[str, str]) -> tuple[str, str, str, bool] | None:
    entrance = tags.get("entrance", "").lower()
    routing_entrance = tags.get("routing:entrance", "").lower()
    barrier = tags.get("barrier", "").lower()
    if (entrance not in USEFUL_ENTRANCES
            and routing_entrance not in {"yes", "main"}
            and barrier not in {"gate", "entrance"}):
        return None
    if tags.get("access", "").lower() in {"no", "private"}:
        return None

    labels = {
        "main": ("المدخل الرئيسي", "Main entrance", "main", True),
        "secondary": ("مدخل ثانوي", "Secondary entrance", "secondary", False),
        "service": ("مدخل الخدمات", "Service entrance", "service", False),
        "emergency": ("مدخل الطوارئ", "Emergency entrance", "emergency", False),
        "parking": ("مدخل المواقف", "Parking entrance", "parking", False),
    }
    if entrance in labels:
        fallback = labels[entrance]
    elif routing_entrance in {"yes", "main"}:
        fallback = ("مدخل", "Entrance", "routing", False)
    else:
        fallback = ("بوابة" if barrier == "gate" else "مدخل", "Gate" if barrier == "gate" else "Entrance", "gate" if barrier == "gate" else "entrance", False)
    name_ar = tags.get("name:ar") or fallback[0]
    name_en = tags.get("name:en") or tags.get("name") or fallback[1]
    return name_ar, name_en, fallback[2], fallback[3]


def main() -> int:
    parser = argparse.ArgumentParser(description="Import source-mapped OSM entrance nodes.")
    parser.add_argument("--pbf", required=True, help="The current merged OSM PBF used by the map updater")
    parser.add_argument("--dry-run", action="store_true", help="Extract and report exact matches without changing the live table")
    args = parser.parse_args()
    pbf_path = Path(args.pbf)
    if not pbf_path.is_absolute():
        pbf_path = APP_DIR / pbf_path
    if not pbf_path.is_file():
        parser.error(f"OSM source file does not exist: {pbf_path}")

    run(["osmium", "--version"])
    print(f"Extracting tagged OSM entrances from {pbf_path} …", flush=True)
    with tempfile.TemporaryDirectory(prefix="siro-osm-gates-") as temp:
        extracted = Path(temp) / "entrances.osm.pbf"
        geojson_path = Path(temp) / "entrances.geojson"
        run([
            "osmium", "tags-filter", str(pbf_path),
            "n/entrance", "n/barrier=gate", "n/barrier=entrance",
            "n/routing:entrance", "-o", str(extracted),
        ])
        run(["osmium", "export", "-u", "type_id", str(extracted), "-o", str(geojson_path)])

        features = json.loads(geojson_path.read_text(encoding="utf-8")).get("features", [])
        rows: list[tuple] = []
        tag_counts: Counter[str] = Counter()
        for feature in features:
            osm_node_id = feature.get("id", "")
            match = re.fullmatch(r"n(-?\d+)", str(osm_node_id))
            geometry = feature.get("geometry") or {}
            coordinates = geometry.get("coordinates") or []
            tags = feature.get("properties") or {}
            if not match or len(coordinates) < 2:
                continue
            classified = label_for(tags)
            if classified is None:
                continue
            lon, lat = coordinates[:2]
            rows.append((
                int(match.group(1)), float(lon), float(lat),
                tags.get("entrance"), tags.get("routing:entrance"), tags.get("barrier"),
            tags.get("access"), tags.get("name"), tags.get("name:ar"),
                tags.get("name:en"), *classified,
                json.dumps(tags, ensure_ascii=False, separators=(",", ":")),
            ))
            tag_counts[tags.get("entrance") or tags.get("barrier") or "other"] += 1

    print(f"Source nodes selected: {len(rows):,}", flush=True)
    if not rows:
        raise RuntimeError("No usable entrance nodes found; keeping the live gate table unchanged.")

    # Stage exact WGS84 source points. This table is not read by the API.
    psql("""
        CREATE TABLE IF NOT EXISTS place_gates_osm_stage (
            osm_id bigint PRIMARY KEY, longitude double precision NOT NULL,
            latitude double precision NOT NULL, entrance text, routing_entrance text,
            barrier text, access text, name text, name_ar text, name_en text,
            gate_name_ar text NOT NULL, gate_name_en text NOT NULL,
            gate_type text NOT NULL, is_main_gate boolean NOT NULL,
            source_tags jsonb NOT NULL
        );
        TRUNCATE place_gates_osm_stage;
    """)
    with tempfile.TemporaryFile(mode="w+t", encoding="utf-8", newline="") as tsv:
        writer = csv.writer(tsv, delimiter="\t", lineterminator="\n", quoting=csv.QUOTE_MINIMAL)
        writer.writerows(rows)
        tsv.seek(0)
        run([
            "docker", "compose", "exec", "-T", "db", "psql", "-X", "-v",
            "ON_ERROR_STOP=1", "-U", os.getenv("DB_USER", "mapuser"), "-d",
            os.getenv("DB_NAME", "mapdb"), "-c",
            "COPY place_gates_osm_stage (osm_id,longitude,latitude,entrance,routing_entrance,barrier,access,name,name_ar,name_en,gate_name_ar,gate_name_en,gate_type,is_main_gate,source_tags) FROM STDIN WITH (FORMAT csv, DELIMITER E'\\t', NULL '')",
        ], input_text=tsv.read())

    psql("""
        ALTER TABLE place_gates_osm_stage ADD COLUMN IF NOT EXISTS geom geometry(Point,3857);
        UPDATE place_gates_osm_stage
        SET geom=ST_Transform(ST_SetSRID(ST_MakePoint(longitude,latitude),4326),3857);
        CREATE INDEX IF NOT EXISTS place_gates_osm_stage_geom_idx
          ON place_gates_osm_stage USING gist (geom);
        ANALYZE place_gates_osm_stage;
    """)

    print("Matching source nodes to OSM facility polygons by exact polygon coverage …", flush=True)
    report_sql = """
        WITH facilities AS (
          SELECT p.osm_id, p.way, r.country_code,
                 'regional_' || r.country_code || '_osm_polygon_' || r.source_id AS place_id
          FROM planet_osm_polygon p
          JOIN regional_place_names r
            ON r.source = 'osm_polygon' AND r.source_id = p.osm_id::text
          WHERE p.name IS NOT NULL
            AND (p.amenity IN ('hospital','clinic','townhall','courthouse')
                 OR p.shop = 'mall' OR p.office = 'government' OR p.building = 'government')
        ), matched AS (
          SELECT g.osm_id AS gate_osm_id, f.country_code, f.place_id
          FROM place_gates_osm_stage g
          JOIN facilities f
            ON ST_Covers(f.way, g.geom)
        ), unique_matches AS (
          SELECT gate_osm_id, min(country_code) AS country_code, min(place_id) AS place_id
          FROM matched GROUP BY gate_osm_id HAVING count(DISTINCT place_id) = 1
        )
        SELECT country_code || '|' || count(*) || '|' || count(DISTINCT place_id)
        FROM unique_matches GROUP BY country_code ORDER BY country_code;
    """
    report = psql(report_sql).strip().splitlines()
    for line in report:
        code, gates, places = line.split("|")
        print(f"{COUNTRY_NAMES.get(code, code)} ({code}): {gates} مصدرًا موثقًا ↔ {places} منشأة", flush=True)
    if not report:
        raise RuntimeError("No source nodes matched named target facilities; the live gate table is unchanged.")

    samples = psql("""
        WITH matches AS (
          SELECT g.osm_id,g.gate_name_ar,g.gate_type,r.country_code,r.name,
                 'regional_'||r.country_code||'_osm_polygon_'||r.source_id AS place_id
          FROM place_gates_osm_stage g
          JOIN planet_osm_polygon p
            ON ST_Covers(p.way,g.geom)
          JOIN regional_place_names r ON r.source='osm_polygon' AND r.source_id=p.osm_id::text
          WHERE p.name IS NOT NULL
            AND (p.amenity IN ('hospital','clinic','townhall','courthouse') OR p.shop='mall'
                 OR p.office='government' OR p.building='government')
        ), unique_ids AS (
          SELECT osm_id,min(place_id) AS place_id FROM matches
          GROUP BY osm_id HAVING count(DISTINCT place_id)=1
        )
        SELECT m.country_code||'|'||m.name||'|'||m.gate_name_ar||'|'||m.gate_type||'|'||m.osm_id
        FROM matches m JOIN unique_ids u ON u.osm_id=m.osm_id AND u.place_id=m.place_id
        ORDER BY m.country_code,m.name LIMIT 12;
    """).strip()
    if samples:
        print("Source-matched examples (country|facility|gate label|OSM tag type|node id):", flush=True)
        print(samples, flush=True)
    if args.dry_run:
        print("Dry run complete; live place_gates was not changed.", flush=True)
        return 0

    # Archive unverifiable legacy rows once, then expose only source-backed OSM rows.
    sync_sql = """
        BEGIN;
        ALTER TABLE place_gates ADD COLUMN IF NOT EXISTS source text;
        ALTER TABLE place_gates ADD COLUMN IF NOT EXISTS source_id text;
        ALTER TABLE place_gates ADD COLUMN IF NOT EXISTS gate_type text;
        ALTER TABLE place_gates ADD COLUMN IF NOT EXISTS access text;
        ALTER TABLE place_gates ADD COLUMN IF NOT EXISTS source_tags jsonb;
        ALTER TABLE place_gates ADD COLUMN IF NOT EXISTS last_seen_at timestamptz;
        CREATE TABLE IF NOT EXISTS place_gates_unverified_archive_20260929 AS
          SELECT * FROM place_gates WHERE source IS NULL AND false;
        INSERT INTO place_gates_unverified_archive_20260929
          SELECT * FROM place_gates WHERE source IS NULL;
        DELETE FROM place_gates WHERE source IS NULL OR source = 'openstreetmap';

        WITH facilities AS (
          SELECT p.osm_id, p.way, r.country_code,
                 'regional_' || r.country_code || '_osm_polygon_' || r.source_id AS place_id
          FROM planet_osm_polygon p
          JOIN regional_place_names r
            ON r.source = 'osm_polygon' AND r.source_id = p.osm_id::text
          WHERE p.name IS NOT NULL
            AND (p.amenity IN ('hospital','clinic','townhall','courthouse')
                 OR p.shop = 'mall' OR p.office = 'government' OR p.building = 'government')
        ), matches AS (
          SELECT g.*, f.place_id
          FROM place_gates_osm_stage g
          JOIN facilities f
            ON ST_Covers(f.way, g.geom)
        ), unique_matches AS (
          SELECT osm_id, min(place_id) AS place_id FROM matches
          GROUP BY osm_id HAVING count(DISTINCT place_id) = 1
        ), exact_matches AS (
          SELECT DISTINCT ON (m.osm_id) m.*
          FROM matches m JOIN unique_matches u ON u.osm_id=m.osm_id AND u.place_id=m.place_id
        )
        INSERT INTO place_gates
          (place_id,gate_name_ar,gate_name_en,latitude,longitude,is_main_gate,source,source_id,gate_type,access,source_tags,last_seen_at)
        SELECT place_id,gate_name_ar,gate_name_en,latitude,longitude,is_main_gate,
               'openstreetmap',osm_id::text,gate_type,access,source_tags,now()
        FROM exact_matches;
        CREATE INDEX IF NOT EXISTS idx_place_gates_source_id ON place_gates(source,source_id);
        CREATE UNIQUE INDEX IF NOT EXISTS idx_place_gates_osm_identity
          ON place_gates(source,source_id,place_id) WHERE source='openstreetmap';
        ANALYZE place_gates;
        COMMIT;
    """
    psql(sync_sql)
    totals = psql("SELECT source || '|' || count(*) FROM place_gates GROUP BY source ORDER BY source;")
    print("Live place_gates rows (source|count):", flush=True)
    print(totals.strip(), flush=True)
    return 0


if __name__ == "__main__":
    try:
        raise SystemExit(main())
    except Exception as exc:  # preserve the existing live table on any failed preflight/import
        print(f"ERROR: {exc}", file=sys.stderr)
        raise SystemExit(1)