#!/bin/bash
# The visitors digest: who is at the table right now and who came today,
# the rooms opened and how far each got, with today's traffic so far.
# Runs ON the droplet:  visitors.sh [YYYY-MM-DD]   (default: today, UTC)
# Single-player duels live in players' browsers, so they show only as
# visits to /play; duels between people show in full from their ledgers.
set -u
DAY="${1:-$(date -u +%F)}"
python3 - "$DAY" <<'PY'
import datetime, glob, json, re, subprocess, sys
from collections import Counter

day = sys.argv[1]
now = datetime.datetime.now(datetime.timezone.utc)
print("as of %s UTC, day %s" % (now.strftime("%H:%M"), day))

def sh(cmd):
    try:
        return subprocess.run(cmd, shell=True, capture_output=True, text=True, timeout=20).stdout.strip()
    except Exception:
        return "?"

print("live sockets: %s | rooms touched in the last hour: %s" % (
    sh("ss -tn state established '( sport = :8788 )' | tail -n +2 | wc -l"),
    sh("find /var/lib/waving-hands/rooms -name '*.jsonl' -mmin -60 2>/dev/null | wc -l")))

# --- traffic: Caddy's access log, people and bots apart ---
BOT = re.compile(r"bot|crawl|spider|slurp|facebookexternalhit|preview|fetch|curl|python|go-http|headless|scan|monitor|uptime", re.I)
PAGES = {"/": "hall", "/play": "play", "/rules": "rules"}
addresses, bots = set(), set()
pages, joins, referrers, campaigns, agents = Counter(), Counter(), Counter(), Counter(), Counter()
requests = 0
for path in sorted(glob.glob("/var/lib/caddy/access.log*")):
    try:
        with open(path) as f:
            for line in f:
                try:
                    x = json.loads(line)
                except Exception:
                    continue
                ts = datetime.datetime.fromtimestamp(x["ts"], datetime.timezone.utc)
                if ts.strftime("%F") != day:
                    continue
                req = x["request"]
                ua = " ".join(req.get("headers", {}).get("User-Agent", [""]))
                ip = req.get("remote_ip") or req.get("client_ip", "?")
                uri = req.get("uri", "")
                path_only = uri.split("?")[0]
                if BOT.search(ua) or not ua:
                    bots.add(ip)
                    continue
                if path_only.startswith("/_app/") or path_only.startswith("/api/") or path_only == "/ws" or path_only.endswith((".png", ".ico", ".txt")):
                    continue
                requests += 1
                addresses.add(ip)
                if path_only in PAGES:
                    pages[PAGES[path_only]] += 1
                elif path_only.startswith("/join/"):
                    joins[path_only.split("/")[2].upper()] += 1
                ref = " ".join(req.get("headers", {}).get("Referer", [""]))
                if ref and "hands.kestrelsnest.social" not in ref:
                    referrers[re.sub(r"^https?://", "", ref).split("/")[0]] += 1
                m = re.search(r"[?&](ref|utm_source|fbclid)=([^&]*)", uri)
                if m:
                    campaigns[m.group(1) + "=" + m.group(2)[:24]] += 1
                agents[re.sub(r"\(.*?\)", "", ua)[:40].strip()] += 1
    except OSError:
        continue
print("traffic: %d addresses of people, %d page requests (%d bot addresses); hall %d, play %d, rules %d, room links %s; referrers %s; campaigns %s" % (
    len(addresses), requests, len(bots), pages["hall"], pages["play"], pages["rules"],
    dict(joins) or "none", dict(referrers) or "none", dict(campaigns) or "none"))

# --- rooms: every ledger, today's in detail ---
first_seen = {}
rooms = []
for f in glob.glob("/var/lib/waving-hands/rooms/*.jsonl"):
    try:
        lines = [json.loads(l) for l in open(f) if l.strip()]
        meta = lines[0]
        assert meta["t"] == "room"
    except Exception:
        continue
    created = datetime.datetime.fromtimestamp(meta["createdAt"] / 1000, datetime.timezone.utc)
    humans, bots_seated, started, turns, last = [], [], False, 0, created
    for x in lines[1:]:
        if x["t"] == "seat":
            (bots_seated if x["bot"] else humans).append(x["name"])
        elif x["t"] == "start":
            started = True
        elif x["t"] == "turn":
            turns += 1
            last = datetime.datetime.fromtimestamp(x["at"] / 1000, datetime.timezone.utc)
    for h in humans:
        if h not in first_seen or created < first_seen[h]:
            first_seen[h] = created
    rooms.append(dict(id=meta["id"], size=meta["seats"], created=created, humans=humans, bots=bots_seated, started=started, turns=turns, last=last))

today = sorted([r for r in rooms if r["created"].strftime("%F") == day], key=lambda r: r["created"])
names = sorted({h for r in today for h in r["humans"]})
new = [n for n in names if first_seen[n].strftime("%F") == day]
print("players today: %s | NEW today: %s" % (", ".join(names) or "none", ", ".join(new) or "none"))
print("rooms today: %d (of %d ever)" % (len(today), len(rooms)))
for r in today:
    state = "lobby only" if not r["started"] else ("started, no turns" if not r["turns"] else "in play")
    span = ("%d min" % round((r["last"] - r["created"]).total_seconds() / 60)) if r["turns"] else ""
    seated = "+".join(r["humans"]) + ((" vs %d bot%s" % (len(r["bots"]), "" if len(r["bots"]) == 1 else "s")) if r["bots"] else "")
    empty = r["size"] - len(r["humans"]) - len(r["bots"])
    print("  %s %s  %s%s  %s  %d turns  %s  last turn %s" % (
        r["created"].strftime("%H:%M"), r["id"], seated, (" (%d seat%s empty)" % (empty, "" if empty == 1 else "s")) if empty else "",
        state, r["turns"], span, r["last"].strftime("%H:%M") if r["turns"] else "never"))
PY