#!/usr/bin/env python3
# ──────────────────────────────────────────────────────────────────────────────
# hns_watchdog.py — PGŽ Sport HNS pipeline watchdog (SUB7)
# Author : Damir Radulić  <dradulic@outlook.com> / <damir@rinet.one>
# Date   : 2026-05-05
# Version: 1.0.0
# Purpose: Periodically poll DB progress for the HNS scraping pipeline,
#          detect stalls, restart fallen worker processes and send Telegram
#          status updates every 30 minutes. Fires a special "ALL DONE" alert
#          once the mission goal is reached.
#
# Modes  : --once    run a single check and exit (cron-friendly)
#          --daemon  loop forever, sleeping CHECK_INTERVAL_SEC between checks
#
# Goal   : 59/59 PGŽ financirani klubovi sa hns_klub_id, ≥80% igrača s
#          profile_complete=true (visina_cm IS NOT NULL), ≥1000 utakmica.
# ──────────────────────────────────────────────────────────────────────────────
import os
import sys
import time
import json
import argparse
import logging
import logging.handlers
import subprocess
from datetime import datetime, timedelta
from pathlib import Path

import psycopg2
import requests

# ── Config ────────────────────────────────────────────────────────────────────
DSN = os.getenv(
    "RINET_DSN",
    f"host=10.10.0.2 port=6432 dbname=rinet_v3 user=rinet password={os.environ['DB_PASSWORD']}",
)
TG_TOKEN = os.getenv("TG_BOT_TOKEN", "8535797835:AAFItT-92jzZ9NWFafLxn0dLa1_n2s-JE5Y")
TG_CHAT  = os.getenv("TG_CHAT", "7969491558")

LOG_DIR  = Path("/var/log/pgz-sport-debug")
LOG_FILE = LOG_DIR / "hns_watchdog.log"
STATE_FILE = LOG_DIR / "hns_watchdog_state.json"

CHECK_INTERVAL_SEC = 30 * 60          # 30 min between daemon iterations
STALL_WINDOW_SEC   = 30 * 60          # consider stale if no growth in 30 min
DONE_FLAG_FILE     = LOG_DIR / "hns_watchdog_DONE.flag"

# Mission targets
TARGET_KLUBOVI       = 59
TARGET_PROFILE_PCT   = 0.80
TARGET_MATCHES       = 1000

# Worker processes to keep alive (process_name : restart_command)
WORKERS = {
    "hns_master_harvester": [
        "python3", "/opt/pgz-sport/scripts/hns_master_harvester.py",
    ],
    "hns_season_v3": [
        "python3", "/opt/pgz-sport/scripts/hns_season_v3.py",
    ],
}

# ── Logging ───────────────────────────────────────────────────────────────────
LOG_DIR.mkdir(parents=True, exist_ok=True)
logger = logging.getLogger("hns_watchdog")
logger.setLevel(logging.INFO)
if not logger.handlers:
    handler = logging.handlers.RotatingFileHandler(
        LOG_FILE, maxBytes=5_000_000, backupCount=5
    )
    handler.setFormatter(logging.Formatter(
        "%(asctime)s %(levelname)s %(message)s",
        datefmt="%Y-%m-%dT%H:%M:%S",
    ))
    logger.addHandler(handler)
    logger.addHandler(logging.StreamHandler(sys.stdout))


# ── Helpers ───────────────────────────────────────────────────────────────────
PROGRESS_SQL = """
SELECT
  (SELECT COUNT(*) FROM pgz_sport.klubovi
     WHERE sport='nogomet' AND pgz_sufinanciran=true
       AND hns_klub_id IS NOT NULL)                                AS klubovi_hns,
  (SELECT COUNT(DISTINCT klub_id) FROM pgz_sport.hns_klub_roster)   AS roster_klubovi,
  (SELECT COUNT(*) FROM pgz_sport.clanovi
     WHERE hns_igrac_id IS NOT NULL)                                AS igraci_hns,
  (SELECT COUNT(*) FROM pgz_sport.clanovi
     WHERE hns_igrac_id IS NOT NULL AND visina_cm IS NOT NULL)      AS igraci_profil,
  (SELECT COUNT(*) FROM pgz_sport.hns_player_seasons)               AS seasons_rec,
  (SELECT COUNT(*) FROM pgz_sport.hns_player_matches)               AS matches_rec
;
"""

PENDING_SQL = """
SELECT COUNT(*) FROM pgz_sport.clanovi
WHERE hns_igrac_id IS NOT NULL
  AND visina_cm IS NULL;
"""


def db_query():
    """Returns dict of progress counters (or None on failure)."""
    try:
        conn = psycopg2.connect(DSN, connect_timeout=10)
        conn.autocommit = True
        with conn.cursor() as cur:
            cur.execute(PROGRESS_SQL)
            row = cur.fetchone()
            cols = ["klubovi_hns", "roster_klubovi", "igraci_hns",
                    "igraci_profil", "seasons_rec", "matches_rec"]
            counts = dict(zip(cols, row))
            try:
                cur.execute(PENDING_SQL)
                counts["pending_players"] = cur.fetchone()[0]
            except Exception as e:
                logger.warning("PENDING_SQL failed: %s", e)
                counts["pending_players"] = None
        conn.close()
        return counts
    except Exception as e:
        logger.error("DB query failed: %s", e)
        return None


def telegram(text):
    try:
        r = requests.post(
            f"https://api.telegram.org/bot{TG_TOKEN}/sendMessage",
            data={"chat_id": TG_CHAT, "text": text[:4000],
                  "parse_mode": "HTML", "disable_web_page_preview": "true"},
            timeout=10,
        )
        ok = r.ok and r.json().get("ok", False)
        if not ok:
            logger.warning("Telegram returned: %s", r.text[:300])
        return ok
    except Exception as e:
        logger.error("Telegram send failed: %s", e)
        return False


def load_state():
    if STATE_FILE.exists():
        try:
            return json.loads(STATE_FILE.read_text())
        except Exception:
            pass
    return {}


def save_state(state):
    try:
        STATE_FILE.write_text(json.dumps(state, indent=2, default=str))
    except Exception as e:
        logger.warning("Cannot persist state: %s", e)


def proc_alive(name):
    """True if a process matching `name` is currently running."""
    try:
        # pgrep -f returns 0 if at least one match
        r = subprocess.run(
            ["pgrep", "-f", name],
            stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL,
            timeout=5,
        )
        return r.returncode == 0
    except Exception as e:
        logger.warning("pgrep failed for %s: %s", name, e)
        return True  # err on caution: do not respawn if uncertain


def restart_worker(name, cmd):
    log_path = LOG_DIR / f"{name}_respawn_{datetime.now():%Y%m%d_%H%M}.log"
    try:
        with open(log_path, "ab") as logf:
            subprocess.Popen(
                cmd,
                stdout=logf, stderr=subprocess.STDOUT,
                cwd="/opt/pgz-sport/scripts",
                start_new_session=True,
            )
        logger.info("Re-spawned worker %s -> %s", name, log_path)
        return True
    except Exception as e:
        logger.error("Failed to respawn %s: %s", name, e)
        return False


def check_workers():
    """Return list of worker names that were re-spawned."""
    respawned = []
    for name, cmd in WORKERS.items():
        if not proc_alive(name):
            if restart_worker(name, cmd):
                respawned.append(name)
    return respawned


def detect_stale(prev, curr):
    """True if seasons_rec did not grow even though there are pending players."""
    if not prev or not curr:
        return False
    if curr.get("pending_players") in (None, 0):
        return False
    try:
        ts_prev = datetime.fromisoformat(prev.get("ts"))
    except Exception:
        return False
    if datetime.utcnow() - ts_prev < timedelta(seconds=STALL_WINDOW_SEC):
        return False  # not enough time elapsed
    grew = (curr.get("seasons_rec", 0) > prev.get("seasons_rec", 0) or
            curr.get("matches_rec", 0) > prev.get("matches_rec", 0) or
            curr.get("igraci_profil", 0) > prev.get("igraci_profil", 0))
    return not grew


def goal_reached(c):
    if not c:
        return False
    if c["klubovi_hns"] < TARGET_KLUBOVI:
        return False
    if c["matches_rec"] < TARGET_MATCHES:
        return False
    if c["igraci_hns"] <= 0:
        return False
    pct = c["igraci_profil"] / c["igraci_hns"]
    return pct >= TARGET_PROFILE_PCT


def fmt_status(c, respawned, stale, suffix=""):
    if not c:
        return f"<b>HNS watchdog</b>\nDB query failed at {datetime.utcnow():%Y-%m-%d %H:%M}Z"
    pct = (c["igraci_profil"] / c["igraci_hns"] * 100) if c["igraci_hns"] else 0
    body = (
        f"<b>HNS watchdog</b> {datetime.utcnow():%Y-%m-%d %H:%MZ}\n"
        f"Klubovi (HNS id):  <b>{c['klubovi_hns']}/{TARGET_KLUBOVI}</b>\n"
        f"Roster scraped:    {c['roster_klubovi']}\n"
        f"Igrači (HNS id):   {c['igraci_hns']}\n"
        f"Igrači s profilom: {c['igraci_profil']} ({pct:0.1f}%)\n"
        f"Sezone:            {c['seasons_rec']}\n"
        f"Utakmice:          <b>{c['matches_rec']}</b>/{TARGET_MATCHES}\n"
        f"Pending igrači:    {c.get('pending_players')}\n"
    )
    if respawned:
        body += f"\nRe-spawned: {', '.join(respawned)}"
    if stale:
        body += "\nSTALE: nema rasta u zadnjih 30 min"
    if suffix:
        body += f"\n{suffix}"
    return body


# ── Main check ────────────────────────────────────────────────────────────────
def run_check(send_telegram=True):
    logger.info("=== watchdog cycle ===")
    state = load_state()
    prev = state.get("last_counts")

    counts = db_query()
    respawned = check_workers()
    stale = detect_stale(prev, counts) if counts else False

    done = goal_reached(counts)
    msg = fmt_status(counts, respawned, stale)

    notify = False
    suffix = ""

    if done and not DONE_FLAG_FILE.exists():
        DONE_FLAG_FILE.write_text(datetime.utcnow().isoformat())
        suffix = "\nALL DONE — mission target reached!"
        msg = fmt_status(counts, respawned, stale, suffix=suffix)
        notify = True
    elif respawned or stale:
        notify = True
    else:
        # routine 30-min heartbeat: send only if last notify >= 30 min ago
        last_ts = state.get("last_notify_ts")
        if not last_ts:
            notify = True
        else:
            try:
                last_dt = datetime.fromisoformat(last_ts)
                if datetime.utcnow() - last_dt >= timedelta(minutes=29):
                    notify = True
            except Exception:
                notify = True

    logger.info("counts=%s respawned=%s stale=%s notify=%s done=%s",
                counts, respawned, stale, notify, done)

    if send_telegram and notify:
        if telegram(msg):
            state["last_notify_ts"] = datetime.utcnow().isoformat()
        else:
            logger.warning("Telegram delivery failed")

    if counts:
        state["last_counts"] = {**counts, "ts": datetime.utcnow().isoformat()}
    save_state(state)

    return {"counts": counts, "respawned": respawned,
            "stale": stale, "done": done, "notified": notify}


# ── Daemon loop ───────────────────────────────────────────────────────────────
def run_daemon():
    logger.info("Starting watchdog daemon (interval=%ss)", CHECK_INTERVAL_SEC)
    while True:
        try:
            run_check(send_telegram=True)
        except Exception as e:
            logger.exception("cycle crashed: %s", e)
        time.sleep(CHECK_INTERVAL_SEC)


# ── Entry point ───────────────────────────────────────────────────────────────
def main():
    p = argparse.ArgumentParser(description="HNS pipeline watchdog")
    g = p.add_mutually_exclusive_group(required=True)
    g.add_argument("--once", action="store_true",
                   help="Run a single check and exit (cron-friendly)")
    g.add_argument("--daemon", action="store_true",
                   help="Run forever, sleeping 30 min between checks")
    p.add_argument("--no-telegram", action="store_true",
                   help="Skip Telegram notifications (debug)")
    args = p.parse_args()

    if args.daemon:
        run_daemon()
    else:
        result = run_check(send_telegram=not args.no_telegram)
        # Print compact JSON for cron / shell usage
        print(json.dumps(result, default=str, ensure_ascii=False))


if __name__ == "__main__":
    main()