#!/usr/bin/env python3
"""Cek hipotesis: jadwal 'anak reply' menyimpan id INDUK sebagai dict char-demi-char di field replies."""
import json, collections
docs = json.load(open('/tmp/threads_reply_dump.json'))
by_id = {d.get('_id'): d for d in docs}


def induk(d):
    for r in d.get("replies") or []:
        if isinstance(r, dict):
            k = sorted(r.keys(), key=lambda x: int(x) if x.isdigit() else 99)
            if len(k) >= 8 and all(x.isdigit() for x in k):
                return "".join(str(r[x]) for x in k)
    return None


anak = []
for d in docs:
    p = induk(d)
    if p:
        anak.append((d.get("_id"), p, p in by_id, (d.get("description") or "")[:70].replace("\n", " ")))
print(f"jadwal dengan penanda induk: {len(anak)}")
for a, p, ada, t in anak[:15]:
    print(f"  anak={a} induk={p} induk_ada_di_dump={ada} | {t}")
# berapa per akun
c = collections.Counter()
for a, p, ada, t in anak:
    c[by_id.get(a, {}).get("accountId") if a in by_id else "?"] += 1
print("\nper akun:", dict(c))
