Files
Jyotisha/scripts/research/reported_offset_sweep.py
T
ClaudeandClaude Opus 5.5 7827d9aff0 chore(rectification): re-freeze after the BUG-1622..1626 renumber (ERR-110)
The renumber (093da446) changed comments in shadbala.py and both replay
scripts, so the frozen identity moved. On top of staging 849343cd:
FREEZE / REPORT constants first, --freeze, then replay:
astrologer_166d_2026_11_05; contract rotated with previous =
astrologer_166c_2026_11_05 (kept byte-for-byte).

Versus astrologer_166c_2026_11_05 and astrologer_166b_2026_11_04: sealed 20
and reported offset 900 all identical, metrics and summary identical,
historical changed_trial_count 10 / 731. Still scoring-15; no tuning.
BUG-1622 records the second re-freeze.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DbBXoyeZrYn6ryNcSe5Y2Y
2026-11-05 15:35:31 +08:00

246 lines
15 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
#!/usr/bin/env python3
"""Reported-centre sensitivity on exposed public cases, without oracle answers."""
from __future__ import annotations
import argparse
import json
import sys
from datetime import datetime, timedelta, timezone
from pathlib import Path
from typing import Any
ROOT = Path(__file__).resolve().parents[2]
if str(ROOT) not in sys.path:
sys.path.insert(0, str(ROOT))
from scripts.active_rectification_event_engine import (
AYANAMSA, NODE_MODE, compute_candidate_static_contexts,
)
from scripts.minute_rectification_blind_eval import _opaque_winner, implementation_sha256
from scripts.minute_rectification_holdout_validator import validate
from scripts.rectification.candidate_contrast import select_signature_representatives
from scripts.rectification.scoring_service import build_event_contribution_matrix, score_from_matrix
from scripts.research.cluster_width_lib import SEPARATION_LEAD, still_valid_public
from scripts.research.probe_supply_after_six import request_from_case
from scripts.research.sealed_holdout_rerun import (
DATASET, file_sha256, historical_comparison, implementation_identity,
opaque_order, verify_frozen_record,
)
OFFSETS = (-30, -20, -15, -10, -8, -5, -3, 0, 3, 5, 8, 10, 15, 20, 30)
RADII = (15, 30, 60)
MINUTE_STEP = 1
# 原值: docs/research/reported_offset_upstream_sync8_2026_10_18.{freeze.json,json}
# 新值: docs/research/reported_offset_upstream_sync9b_2026_10_28.{freeze.json,json}
# 原因: sync9b(BUG-1494~1495)改了 scripts/varga.py,它在扩展身份里;报时窗口输出不变,ERR-110 另存新记录
# 原值: docs/research/reported_offset_upstream_sync9b_2026_10_28.{freeze.json,json}
# 新值: docs/research/reported_offset_d60_mapping_2026_10_29.{freeze.json,json}
# 原因: BUG-1500 修正生产 D60 映射,scripts/varga.py 变了(在扩展身份里);900 条回放不变,ERR-110 另存新记录
# 原值: docs/research/reported_offset_d60_mapping_2026_10_29.{freeze.json,json}
# 新值: docs/research/reported_offset_data_audit_2026_11_04.{freeze.json,json}
# 原因: 数据核对 BUG-1545(narayana_dasha.py、active_rectification_event_engine.py)与 BUG-1546(shadbala.py)改了冻结身份文件,ERR-110 另存新记录
# 原值: docs/research/reported_offset_data_audit_2026_11_04.{freeze.json,json}
# 新值: docs/research/reported_offset_moon_paksha_2026_11_04.{freeze.json,json}
# 原因: BUG-1557 月亮 Paksha 口径(scripts/shadbala.py)改了冻结身份文件,ERR-110 另存新记录
# 原值: docs/research/reported_offset_moon_paksha_2026_11_04.{freeze.json,json}
# 新值: docs/research/reported_offset_audit_wave2_2026_11_04.{freeze.json,json}
# 原因: 数据核对第二波合并(BUG-1560 / BUG-1561 varga.py、divisional_charts_extended.py;BUG-1566 / BUG-1567 jaimini.py;BUG-1270 / BUG-1576 shadbala.py)改了冻结身份文件,ERR-110 在合并后的代码上另存新记录;各分支自己的 audit_g / audit_h / audit_j 记录只对各自分支成立,作为中间记录原样保留
# 原值: docs/research/reported_offset_audit_wave2_2026_11_04.{freeze.json,json}
# 新值: docs/research/reported_offset_d30_venus_2026_11_04.{freeze.json,json}
# 原因: 产品 2026-11-04 决定按出处改算法:BUG-1560 varga.py 偶数星座 D30 按 BPHS 第 6 章、BUG-1570 白羊 / 天蝎上升金星判功能凶(BPHS 34.20 / 34.35),校正身份升到 scoring-13;冻结身份文件变了且回放结果变了,ERR-110 另存新记录
# 原值: docs/research/reported_offset_d30_venus_2026_11_04.{freeze.json,json}
# 新值: docs/research/reported_offset_astrologer_166_2026_11_04.{freeze.json,json}
# 原因: #166 占星师裁定(BUG-1581~1589,产品 2026-11-04 决定交付)改了 shadbala.py、jaimini.py、varga.py、divisional_charts_extended.py,校正身份升到 scoring-14;ERR-110 另存新记录
# 原值: docs/research/reported_offset_astrologer_166_2026_11_04.{freeze.json,json}
# 新值: docs/research/reported_offset_astrologer_166b_2026_11_04.{freeze.json,json}
# 原因: #166 第二批(BUG-1590~1598)改了 functional_benefics.py、jaimini.py、narayana_dasha.py、shadbala.py、varga.py,校正身份升到 scoring-15;ERR-110 另存新记录
# 原值: docs/research/reported_offset_astrologer_166b_2026_11_04.{freeze.json,json}
# 新值: docs/research/reported_offset_astrologer_166c_2026_11_05.{freeze.json,json}
# 原因: #166 占星师 2026-11-05 答复(BUG-1622~1626)只改了 shadbala.py 的标签、注释与共用的水星规则函数(数值不变),校正身份仍是 scoring-15;ERR-110 另存新记录
# 原值: docs/research/reported_offset_astrologer_166c_2026_11_05.{freeze.json,json}
# 新值: docs/research/reported_offset_astrologer_166d_2026_11_05.{freeze.json,json}
# 原因: BUG-1622~1626 改号(另一会话已用 BUG-1613)改了 shadbala.py 与本脚本的注释,冻结身份变了,数值不变,仍 scoring-15;ERR-110 另存新记录
FREEZE = ROOT / "docs/research/reported_offset_astrologer_166d_2026_11_05.freeze.json"
REPORT = ROOT / "docs/research/reported_offset_astrologer_166d_2026_11_05.json"
LEGACY_REPORT = ROOT / "docs/research/reported_offset_2026_09_20.json"
def shifted_window(case: dict[str, Any], offset: int, radius: int) -> tuple[dict[str, Any], list[datetime]]:
if radius not in (*RADII, 120):
raise ValueError("unsupported_radius")
true_at = datetime.fromisoformat(f"{case['birth']['date']}T{case['birth']['time']}:00")
reported_at = true_at + timedelta(minutes=offset)
start = reported_at - timedelta(minutes=radius)
end = reported_at + timedelta(minutes=radius)
candidates = [start + timedelta(minutes=i) for i in range(0, radius * 2 + 1, MINUTE_STEP)]
# Reuse only the established event normalization, then replace the window.
# The ranker sees neither the truth label nor the simulated offset.
request = request_from_case(case)
request.update({
"birth_date": start.date().isoformat(),
"start_time": start.strftime("%H:%M"),
"end_time": end.strftime("%H:%M"),
"minute_step": MINUTE_STEP,
"ayanamsa": AYANAMSA,
"node_mode": NODE_MODE,
})
return request, candidates
def score_window(request: dict[str, Any], contexts: list[dict[str, Any]]) -> list[dict[str, Any]]:
"""Use the repaired native single matrix, with candidate dates preserved.
The historical grouped adapter is archived, not a second active scoring
path. This is a native scoring replay, still not a production Q&A replay.
"""
built = build_event_contribution_matrix(request, static_contexts=contexts)
return score_from_matrix(request, built)
def delivery_moments(public: list[dict[str, Any]], candidates: list[datetime]) -> list[datetime]:
"""Initial delivery envelope in date-aware order; no truth-based narrowing."""
valid = still_valid_public(public, {}, lead=SEPARATION_LEAD)
clocks = {str(value)[:5] for row in valid for value in (row.get("cluster_times") or [row["time"]])}
return [candidate for candidate in candidates if candidate.strftime("%H:%M") in clocks]
def reveal_metrics(
rows: list[dict[str, Any]], candidates: list[datetime], delivery: list[datetime],
truth: datetime, benchmark_id: str, case_id: str,
) -> dict[str, Any]:
ordered = opaque_order(benchmark_id, case_id, rows)
predicted_clock = _opaque_winner(benchmark_id, case_id, rows)
by_clock = {candidate.strftime("%H:%M"): candidate for candidate in candidates}
if len(by_clock) != len(candidates):
raise ValueError("ambiguous_candidate_clock")
predicted = by_clock[predicted_clock]
within = min(candidates) <= truth <= max(candidates)
in_candidates = truth in candidates
rank = next((i for i, row in enumerate(ordered, 1) if by_clock[row["time"]] == truth), None)
covered = bool(delivery) and min(delivery) <= truth <= max(delivery)
return {
"truth_in_window": within,
"truth_in_candidates": in_candidates,
"true_rank": rank,
"top_1_hit": predicted == truth,
"top_1_minute_error": abs((predicted - truth).total_seconds()) / 60,
"delivery_covers_truth": covered,
"candidate_count": len(candidates),
"delivery_width_minutes": int((max(delivery) - min(delivery)).total_seconds() / 60) + 1 if delivery else 0,
}
def summarize(trials: list[dict[str, Any]], radii: tuple[int, ...], offsets: tuple[int, ...]) -> list[dict[str, Any]]:
result = []
for radius in radii:
for offset in offsets:
group = [row for row in trials if row["radius_minutes"] == radius and row["offset_minutes"] == offset]
count = len(group)
result.append({
"radius_minutes": radius, "offset_minutes": offset, "trial_count": count,
**{key: round(sum(row[source] for row in group) / count, 4) if count else None for key, source in (
("truth_in_window_rate", "truth_in_window"),
("top_1_rate", "top_1_hit"),
("delivery_coverage_rate", "delivery_covers_truth"),
("mean_absolute_minute_error", "top_1_minute_error"),
)},
})
return result
def freeze_record(dataset: Path = DATASET, radii: tuple[int, ...] = RADII, offsets: tuple[int, ...] = OFFSETS) -> dict[str, Any]:
manifest = json.loads(dataset.read_text(encoding="utf-8"))
legacy_hash = implementation_sha256(manifest["frozen_scoring"]["files"])
return {
"record_version": "reported-offset-native-candidate-date-v3",
"frozen_at_utc": datetime.now(timezone.utc).isoformat(),
"ayanamsa": AYANAMSA, "node_mode": NODE_MODE,
"radii_minutes": list(radii), "offsets_minutes": list(offsets), "minute_step": MINUTE_STEP,
"dataset": dataset.relative_to(ROOT).as_posix(), "dataset_sha256": file_sha256(dataset),
"dataset_benchmark_id": manifest["benchmark_id"],
"implementation_sha256": legacy_hash,
"implementation_sha256_prefix": legacy_hash[:16],
**implementation_identity(dataset),
"scorer": "native_event_contribution_matrix_not_shadow_fact_ranker",
"delivery": "initial_signature_clusters_peak_gap_lt_8_envelope_no_answers_no_elimination",
"rank": "score_desc_then_sha256(benchmark_id:case_id:candidate_time)",
"grid": "every_minute_inclusive_not_production_two_minute_sampling",
"cross_midnight": "date_aware_candidates_distance_and_native_single_matrix",
"evaluator_sha256": file_sha256(Path(__file__)),
"replay_revision": "native_candidate_date_v3",
"supersedes": "candidate_date_grouped_v2_identity_and_path_not_assumed_numerically_wrong",
"is_blind_evaluation": False, "truth_hidden_from_ranker": True,
"official_valid_independent_blind": False, "official_blind_trial_count": 0,
"results_previously_seen": True, "must_not_use_for_tuning": True,
}
def run(dataset: Path = DATASET, radii: tuple[int, ...] = RADII, offsets: tuple[int, ...] = OFFSETS,
freeze_path: Path = FREEZE) -> dict[str, Any]:
frozen = json.loads(freeze_path.read_text(encoding="utf-8"))
verify_frozen_record(frozen, freeze_record(dataset, radii, offsets))
replay_started_at = datetime.now(timezone.utc).isoformat()
manifest = json.loads(dataset.read_text(encoding="utf-8"))
validation = validate(dataset)
invalid = validation["invalid_cases"]
trials = []
for index, case in enumerate(manifest["cases"], 1):
if case["case_id"] in invalid:
continue
# Reuse static chart calculations, not window-dependent scores or ranks.
contexts_by_moment: dict[datetime, dict[str, Any]] = {}
for radius in radii:
for offset in offsets:
request, candidates = shifted_window(case, offset, radius)
missing = [moment for moment in candidates if moment not in contexts_by_moment]
if missing:
contexts_by_moment.update(zip(missing, compute_candidate_static_contexts(request, candidates=missing), strict=True))
contexts = [contexts_by_moment[moment] for moment in candidates]
rows = score_window(request, contexts)
public = select_signature_representatives(rows, contexts)
delivery = delivery_moments(public, candidates)
# All scores and the delivery envelope are locked before reveal.
truth = datetime.fromisoformat(f"{case['birth']['date']}T{case['birth']['time']}:00")
trials.append({
"case_ordinal": index, "offset_minutes": offset, "radius_minutes": radius,
**reveal_metrics(rows, candidates, delivery, truth, manifest["benchmark_id"], case["case_id"]),
})
verify_frozen_record(frozen, freeze_record(dataset, radii, offsets))
return {
"scope": "reported_offset_sensitivity_not_product_accuracy",
"replay_started_at_utc": replay_started_at,
"replay_finished_at_utc": datetime.now(timezone.utc).isoformat(),
"specification": frozen,
"frozen_record": frozen,
"freeze_record_path": freeze_path.relative_to(ROOT).as_posix(),
"implementation_hash_matches_at_replay": True,
"dataset_hash_matches_at_replay": True,
"historical_comparison": historical_comparison(LEGACY_REPORT, trials, ("case_ordinal", "radius_minutes", "offset_minutes")),
"excluded_cases": invalid, "case_count": validation["valid_public_aa_cases"],
"trial_count": len(trials), "trials": trials,
"summary": summarize(trials, radii, offsets),
"widening_geometry": {
"scored_radii": list(radii),
"radius_120_scored": 120 in radii,
"truth_in_window_condition": "abs(reported_offset_minutes) <= radius_minutes",
"radius_15_first_integer_minute_outside": 16,
"all_tested_offsets_within_radius_30_60_120": max(map(abs, offsets)) <= 30,
"boundary": "Geometry rescues candidate inclusion only, not ranking or delivery coverage; no population frequency without a reported-offset distribution.",
},
}
if __name__ == "__main__":
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--json", action="store_true")
parser.add_argument("--freeze", action="store_true", help="Exclusively create the record before replay")
parser.add_argument("--freeze-path", type=Path, default=FREEZE)
args = parser.parse_args()
if args.freeze:
with args.freeze_path.open("x", encoding="utf-8", newline="\n") as handle:
json.dump(freeze_record(), handle, ensure_ascii=False, indent=2)
handle.write("\n")
print(args.freeze_path.relative_to(ROOT).as_posix())
else:
print(json.dumps(run(freeze_path=args.freeze_path), ensure_ascii=False, indent=2))