SniperGold_ML/ml/p3/smc_semantic/audit_choch_state.py

# -*- coding: utf-8 -*-
"""P3-S.3 AUDIT — CHOCH STATE/EVENT FORENSIC (diagnostic, mirror f7 P3-S.0).

Mengukur perilaku POPULASI state CHoCH (f8/f9/f18) pada XAUUSD M15 2017-2026
dengan semantik RUNTIME (equal-allowed pivot, build_structure_fast):

  - onsets      : bar di mana CHoCH event BARU tercatat (choch_bar == r)
  - active bars : bar dengan choch_dir != 0 (state f8 "last CHoCH")
  - repetition  : active/onsets
  - run stats   : panjang run state (dir konstan) & age (r - choch_bar)
  - f9 staleness: chochOK (chochBar >= swpBar & searah) tanpa expiry -> berapa
                  banyak bar dengan CHoCH age > SEQ_WINDOW masih "confirm"
  - f18 staleness: kontribusi +15 (chochDir==swpDir) pada CHoCH basi

Output: output/audit_choch_state.json
Disiplin: OBSERVE -> RECORD. Bukan oracle; bukan definisi. Spec = truth.
"""
import os
import sys
import json

import numpy as np

HERE = os.path.dirname(os.path.abspath(__file__))
REPO_ML = os.path.normpath(os.path.join(HERE, "..", "..", "..", "ml"))      # publish/AlgoForge/ml
REPO_P3 = os.path.normpath(os.path.join(REPO_ML, "p3"))
REPO_PARITY = os.path.normpath(os.path.join(REPO_ML, "parity"))
SRC_LEGACY = os.path.normpath(os.path.join(
    HERE, "..", "..", "..", "..", "..", "SniperGold_ML"))                    # Shared Projects\SniperGold_ML
sys.path.insert(0, HERE)
sys.path.insert(0, REPO_ML)
sys.path.insert(0, REPO_P3)
sys.path.insert(0, REPO_PARITY)
sys.path.insert(0, SRC_LEGACY)

import p3_common as P3
import smc_semantic_common as SC
import build_features_p2 as BFP

# Checkout ini nested (publish\AlgoForge) -> path relatif p3_common meleset;
# data riil ada di MQL5\Files\AlgoForge\Data. Patch konstanta data root.
P3.DATA = r"D:\TradingTerminal\HFM Metatrader 5\MQL5\Files\AlgoForge\Data"

SEQ_WINDOW = SC.SEQ_WINDOW   # 40


def main():
    t, o, h, l, c, v, htf = SC.load_data()
    n = len(c)

    # ---- struktur runtime-semantics (equal allowed), full-feed, begin=100 ----
    sw_at = np.zeros(n, dtype=int)
    sw = BFP.build_structure_fast(o, h, l, c, SC.SWING_LEN, False, sw_at, begin=100)
    inn = BFP.build_structure_fast(o, h, l, c, SC.INTERNAL_LEN, True,
                                   sw["trend"].copy(), begin=100)

    cd = inn["choch_dir"]          # state f8 (last CHoCH direction)
    cb = inn["choch_bar"]          # bar event CHoCH terakhir
    new_event = (cb == np.arange(n))   # bar di mana event BARU tercatat

    onsets = int(new_event.sum())
    active = int((cd != 0).sum())
    rep = active / onsets if onsets else 0.0

    # run stat: panjang run arah konstan (cd != 0)
    runs = []
    run_len = 0
    prev = 0
    for x in cd:
        if x != 0 and x == prev:
            run_len += 1
        else:
            if prev != 0:
                runs.append(run_len)
            run_len = 1 if x != 0 else 0
        prev = x
    if prev != 0:
        runs.append(run_len)
    runs = np.array(runs, dtype=int) if runs else np.zeros(0, dtype=int)

    # age state (bar setelah event)
    mask = cb >= 0
    age = np.where(mask, np.arange(n) - cb, -1)
    age_active = age[(cd != 0) & mask]

    # ---- sweep state legacy (utk f9/f18); pivot list = feed internal build_structure_fast ----
    is_ph, is_pl = BFP._pivots_vec(h, l, SC.INTERNAL_LEN)
    piv = []
    for p in range(SC.INTERNAL_LEN, n - SC.INTERNAL_LEN):
        if p >= 95 and (is_ph[p] or is_pl[p]):
            piv.append((p, float(h[p]) if is_ph[p] else float(l[p]), bool(is_ph[p])))
    sweep_dir, sweep_bar = SC.sweep_state(h, l, c, piv)

    # f9 chochOK (kode EA, tanpa expiry): chochDir!=0 & chochBar>=swpBar & dir sama
    chochOK = np.zeros(n, dtype=bool)
    for r in range(n):
        if cd[r] != 0 and cb[r] >= sweep_bar[r] and cd[r] == sweep_dir[r]:
            chochOK[r] = True

    # f9 pada CHoCH basi (age > SEQ_WINDOW)
    ok_basi = chochOK & (age > SEQ_WINDOW)
    ok_fresh = chochOK & (age >= 0) & (age <= SEQ_WINDOW)

    # f18 +15 leg: chochDir==swpDir (tanpa expiry)
    f18_leg = (cd != 0) & (cd == sweep_dir)
    f18_leg_basi = f18_leg & (age > SEQ_WINDOW)

    # kapan state f8 pertama kali menjadi 0 setelah pernah aktif (post-first-event)
    first_active = int(np.argmax(cd != 0)) if (cd != 0).any() else -1
    zero_after_first = bool((cd[first_active:] == 0).any()) if first_active >= 0 else None

    result = {
        "scope": "XAUUSD M15 2017-2026 full-feed (runtime-semantics equal-allowed pivots)",
        "n_bars": int(n),
        "choch_state": {
            "onsets": onsets,
            "active_bars": int(active),
            "repetition_ratio": round(float(rep), 3),
            "zero_after_first_event": zero_after_first,
            "run_median": float(np.median(runs)) if len(runs) else None,
            "run_max": int(runs.max()) if len(runs) else None,
            "run_n": int(len(runs)),
            "age_active_median": float(np.median(age_active)) if len(age_active) else None,
            "age_active_p99": float(np.percentile(age_active, 99)) if len(age_active) else None,
            "age_active_max": int(age_active.max()) if len(age_active) else None,
            "note": "state f8 = last CHoCH direction (persisten, per FEATURE_CONTRACT); "
                    "bukan duplicate emission detector."
        },
        "f9_chochOK": {
            "true_bars": int(chochOK.sum()),
            "fresh_bars_le40": int(ok_fresh.sum()),
            "stale_bars_gt40": int(ok_basi.sum()),
            "stale_fraction": round(float(ok_basi.sum() / max(1, int(chochOK.sum()))), 4),
            "note": "f9 memakai CHoCH tanpa expiry -> event basi tetap 'confirm'"
        },
        "f18_leg_choch_swp": {
            "true_bars": int(f18_leg.sum()),
            "stale_bars_gt40": int(f18_leg_basi.sum()),
            "stale_fraction": round(float(f18_leg_basi.sum() / max(1, int(f18_leg.sum()))), 4),
            "note": "f18 +15 (chochDir==swpDir) tanpa expiry"
        },
        "seq_window": SEQ_WINDOW,
    }

    out_path = os.path.join(HERE, "output", "audit_choch_state.json")
    os.makedirs(os.path.dirname(out_path), exist_ok=True)
    with open(out_path, "w", encoding="utf-8") as f:
        json.dump(result, f, indent=2, default=str)
    print(json.dumps(result, indent=2, default=str))
    print("[saved]", out_path)


if __name__ == "__main__":
    main()