#!/usr/bin/env python3
"""build_gtfs_db.py — GTFS-Datei in SQLite laden (einmalig).
Aufruf: python3 build_gtfs_db.py <gtfs_zip> <out_db> [gtfs_dir]
Legt Tabellen stops/trips/st an (Schema siehe gtfs-timetable-routing/references).
"""
import zipfile, sqlite3, os, sys, time

def main():
    gtfs_zip = sys.argv[1]
    out_db = sys.argv[2]
    work = sys.argv[3] if len(sys.argv) > 3 else "/tmp/gtfs_x"
    os.makedirs(work, exist_ok=True)
    # extrahieren
    with zipfile.ZipFile(gtfs_zip) as z:
        z.extractall(work)
    t0 = time.time()
    con = sqlite3.connect(out_db)
    con.execute("PRAGMA journal_mode=MEMORY"); con.execute("PRAGMA temp_store=MEMORY")
    con.executescript("""
    CREATE TABLE stops(stop_id TEXT PRIMARY KEY, stop_name TEXT, stop_lat REAL, stop_lon REAL);
    CREATE TABLE trips(trip_id TEXT PRIMARY KEY, route_id TEXT, service_id TEXT);
    CREATE TABLE st(stop_id TEXT, trip_id TEXT, seq INT, arr TEXT, dep TEXT);
    CREATE INDEX idx_st_stop ON st(stop_id);
    CREATE INDEX idx_st_trip ON st(trip_id);
    CREATE INDEX idx_st_seq ON st(trip_id, seq);
    """)
    # stops.txt: stop_name,parent_station,stop_id,stop_lat,stop_lon,...
    n=0
    with open(f"{work}/stops.txt", encoding="utf-8") as f:
        f.readline()
        for line in f:
            p = line.strip().split(",")
            if len(p) >= 5 and p[3] and p[4]:
                sid = p[2] if p[2] else p[1]  # bevorzugt stop_id, fallback parent_station
                con.execute("INSERT OR IGNORE INTO stops VALUES(?,?,?,?)", (sid, p[0], p[3], p[4]))
                n += 1
    print("stops:", n)
    # trips.txt: route_id,service_id,trip_id
    n=0
    with open(f"{work}/trips.txt", encoding="utf-8") as f:
        f.readline()
        for line in f:
            p = line.strip().split(",")
            if len(p) >= 3 and p[0]:
                con.execute("INSERT OR IGNORE INTO trips VALUES(?,?,?)", (p[2], p[0], p[1]))
                n += 1
    print("trips:", n)
    # stop_times.txt: trip_id,arrival_time,departure_time,stop_id,stop_sequence,...
    n=0; batch=[]
    with open(f"{work}/stop_times.txt", encoding="utf-8") as f:
        f.readline()
        for line in f:
            p = line.strip().split(",")
            if len(p) >= 5 and p[0] and p[3]:
                batch.append((p[3], p[0], int(p[4]) if p[4].isdigit() else 0, p[1], p[2]))
                if len(batch) >= 50000:
                    con.executemany("INSERT INTO st VALUES(?,?,?,?,?)", batch); n += len(batch); batch = []
        if batch:
            con.executemany("INSERT INTO st VALUES(?,?,?,?,?)", batch); n += len(batch)
    con.commit(); con.close()
    print("stop_times:", n, "| DB gebaut in %.1fs" % (time.time()-t0))

if __name__ == "__main__":
    main()