#!/usr/bin/env bash
# source: /home/_crossfeed/tooling/agent-stack @ HEAD
# agent-metrics.sh — harvest per-vertical crush usage into the org ledger.
# Appends one JSONL line per session seen, deduped by session id via
# ledger/seen index. Idempotent; safe to run every timer tick.
#
# Ledger: /home/_crossfeed/metrics/ledger.jsonl
# Event:  {"ts","vertical","role","session","model","prompt_tokens",
#          "completion_tokens","total_tokens","cost"}
# Role heuristic: title contains 'PMO intro' -> pmo; 'INTRO-WORK' -> work;
# 'headless' -> workturn; else 'other'.
set -uo pipefail

METRICS=${METRICS:-/home/_crossfeed/metrics}
LEDGER="$METRICS/ledger.jsonl"
SEEN="$METRICS/.seen-ids"
VERTICALS=(TSGBOD TSGCTO TSGCCO TSGCOO reachableceo)
RUNAS=${RUNAS:-sudo}   # agent-metrics runs from the supervisor service (root)

mkdir -p "$METRICS"; touch "$LEDGER" "$SEEN"

emit() { printf '%s\n' "$1" >> "$LEDGER"; }

for user in "${VERTICALS[@]}"; do
  home=$(getent passwd "$user" | cut -d: -f6)
  [ -n "$home" ] || continue
  # shellcheck disable=SC2016
  json=$($RUNAS runuser -u "$user" -- env HOME="$home" TERM=xterm-256color \
    bash -c "cd '$home' && crush session list --json 2>/dev/null" || true)
  [ -n "$json" ] || continue
  # one session id per line; token/cost fields only exist in `session show`
  while IFS= read -r sid; do
    [ -n "$sid" ] || continue
    grep -q "$sid" "$SEEN" && continue
    obj=$($RUNAS runuser -u "$user" -- env HOME="$home" TERM=xterm-256color \
      bash -c "cd '$home' && crush session show '$sid' --json 2>/dev/null | head -c 4000" || true)
    [ -n "$obj" ] || continue
    title=$(printf '%s' "$obj" | grep -o '"title":"[^"]*"' | head -1 | cut -d'"' -f4)
    role=other
    case "$title" in
      *PMO*intro*) role=pmo ;;
      *INTRO-WORK*) role=work ;;
      *headless*|*TSG*-Work*executing*) role=workturn ;;
    esac
    model=$(printf '%s' "$obj" | grep -o '"model":"[^"]*"' | head -1 | cut -d'"' -f4)
    pt=$(printf '%s' "$obj" | grep -o '"prompt_tokens":[0-9]*' | head -1 | cut -d: -f2)
    ct=$(printf '%s' "$obj" | grep -o '"completion_tokens":[0-9]*' | head -1 | cut -d: -f2)
    tt=$(printf '%s' "$obj" | grep -o '"total_tokens":[0-9]*' | head -1 | cut -d: -f2)
    cost=$(printf '%s' "$obj" | grep -o '"cost":[0-9.e-]*' | head -1 | cut -d: -f2)
    ts=$(date -Is)
    emit "{\"ts\":\"$ts\",\"vertical\":\"$user\",\"role\":\"$role\",\"session\":\"$sid\",\"model\":\"${model:-?}\",\"prompt_tokens\":${pt:-0},\"completion_tokens\":${ct:-0},\"total_tokens\":${tt:-0},\"cost\":${cost:-0}}"
    echo "$sid" >> "$SEEN"
  done < <(printf '%s' "$json" | grep -o '"id":"[0-9a-f]*"' | cut -d'"' -f4)
done

# rolling summary
python3 - "$LEDGER" <<'PYEOF' > "$METRICS/SUMMARY.md" 2>/dev/null || true
import json, sys, collections, datetime
tot = collections.Counter(); cost = collections.Counter(); n = collections.Counter()
day = collections.Counter()
now = datetime.datetime.now().astimezone()
for line in open(sys.argv[1]):
    try: e = json.loads(line)
    except Exception: continue
    v = e.get("vertical","?"); t = e.get("total_tokens",0)
    tot[v] += t; cost[v] += e.get("cost",0.0); n[v] += 1
    try:
        ts = datetime.datetime.fromisoformat(e["ts"])
        if (now-ts).total_seconds() < 86400: day[v] += t
    except Exception: pass
print(f"# Agent usage summary — {now.isoformat()}")
print(f"- sessions ledgered total: {sum(n.values())}")
print("- all-time by vertical:")
for v in sorted(tot):
    print(f"  {v:<13} n={n[v]:<4} tokens={tot[v]:<10} cost={cost[v]:.4f} (24h tokens={day[v]})")
PYEOF