File size: 9,425 Bytes
2955ecc | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 | #!/usr/bin/env python3
"""validate synthetic observable sites, expected outputs, and demo data"""
from __future__ import annotations
import argparse
import json
import shutil
import subprocess
import sys
from pathlib import Path
from typing import Any
import yaml
ROOT = Path(__file__).resolve().parents[1]
SRC = ROOT / "src"
if str(SRC) not in sys.path:
sys.path.insert(0, str(SRC))
from datacenter_verification.observable_algorithm import evaluate_site
def main() -> None:
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--input", type=Path, default=ROOT / "synthetic" / "sites.json")
parser.add_argument("--website-dir", type=Path)
parser.add_argument("--check-website-js", action="store_true")
args = parser.parse_args()
data = json.loads(args.input.read_text(encoding="utf-8"))
feature_schemas = load_feature_schemas(ROOT / "observables" / "observables.yaml")
source_refs = load_source_refs(ROOT / "observables" / "rules" / "source_ledger.yaml")
errors: list[str] = []
results: dict[str, dict[str, Any]] = {}
for site in data["sites"]:
errors.extend(validate_raw_features(site, feature_schemas))
errors.extend(validate_source_refs(site, source_refs))
result = evaluate_site(site)
results[site["site_id"]] = result
errors.extend(validate_expected(site, result))
errors.extend(validate_stage_invariants(site, result))
if args.website_dir:
errors.extend(validate_website_data(args.website_dir, results))
if args.check_website_js:
errors.extend(validate_website_js(args.website_dir))
if errors:
for error in errors:
print(f"ERROR: {error}", file=sys.stderr)
raise SystemExit(1)
print(
json.dumps(
{
"sites": len(data["sites"]),
"raw_feature_ids_validated": len(feature_schemas),
"website_checked": bool(args.website_dir),
"website_js_checked": bool(args.check_website_js),
},
indent=2,
)
)
def load_feature_schemas(path: Path) -> dict[str, set[str]]:
payload = yaml.safe_load(path.read_text(encoding="utf-8"))
schemas: dict[str, set[str]] = {}
for observable in payload["observables"]:
for feature in observable.get("features", []):
fields = {
field["name"]
for field in feature.get("value_schema", {}).get("fields", [])
if isinstance(field, dict) and "name" in field
}
schemas[feature["id"]] = fields
return schemas
def load_source_refs(path: Path) -> set[str]:
payload = yaml.safe_load(path.read_text(encoding="utf-8"))
refs = set()
for item in payload.get("sources", []):
if "id" in item:
refs.add(item["id"])
return refs
def validate_raw_features(site: dict[str, Any], schemas: dict[str, set[str]]) -> list[str]:
errors = []
for feature_id, records in site.get("raw_features", {}).items():
if feature_id not in schemas:
errors.append(f"{site['site_id']}: raw feature id {feature_id!r} is not in observables/observables.yaml")
continue
if not isinstance(records, list):
errors.append(f"{site['site_id']}: raw feature {feature_id!r} must be a list")
continue
allowed = schemas[feature_id]
for index, record in enumerate(records):
if not isinstance(record, dict):
errors.append(f"{site['site_id']}: raw feature {feature_id}[{index}] must be an object")
continue
unknown = sorted(set(record) - allowed)
if unknown:
errors.append(f"{site['site_id']}: raw feature {feature_id}[{index}] has unknown fields {unknown}")
return errors
def validate_source_refs(site: dict[str, Any], source_refs: set[str]) -> list[str]:
errors = []
for ref in site.get("source_refs", []):
if ref not in source_refs:
errors.append(f"{site['site_id']}: source ref {ref!r} missing from source ledger")
return errors
def validate_expected(site: dict[str, Any], result: dict[str, Any]) -> list[str]:
expected = site["expected"]
stages = result["stage_outputs"]
errors = []
if stages["A_capacity_gate"]["label"] != expected["A_capacity_gate_label"]:
errors.append(
f"{site['site_id']}: A expected {expected['A_capacity_gate_label']} got {stages['A_capacity_gate']['label']}"
)
for label in expected["B_training_candidate_detection_labels"]:
if label not in stages["B_training_candidate_detection"]["labels"]:
errors.append(f"{site['site_id']}: B missing expected label {label}")
if not expected["B_training_candidate_detection_labels"] and stages["B_training_candidate_detection"]["labels"]:
errors.append(f"{site['site_id']}: B expected no labels got {stages['B_training_candidate_detection']['labels']}")
for label in expected["C_discrepancy_and_explanation_review_labels"]:
if label not in stages["C_discrepancy_and_explanation_review"]["labels"]:
errors.append(f"{site['site_id']}: C missing expected label {label}")
if result["final_route"] != expected["final_route"]:
errors.append(f"{site['site_id']}: final expected {expected['final_route']} got {result['final_route']}")
if stages["A_capacity_gate"]["short_circuited"] != expected["capacity_short_circuit"]:
errors.append(f"{site['site_id']}: capacity short-circuit mismatch")
return errors
def validate_stage_invariants(site: dict[str, Any], result: dict[str, Any]) -> list[str]:
errors = []
stages = result["stage_outputs"]
a = stages["A_capacity_gate"]
b = stages["B_training_candidate_detection"]
c = stages["C_discrepancy_and_explanation_review"]
final = result["final_route"]
if a["label"] == "capacity_ruled_out_for_scope":
if b["mode"] != "skipped_due_to_capacity_ruleout" or c["mode"] != "skipped_due_to_capacity_ruleout":
errors.append(f"{site['site_id']}: capacity rule-out did not short-circuit B/C")
if not b["labels"] and a["label"] != "capacity_ruled_out_for_scope":
if c["mode"] != "C1_negative_screen_integrity":
errors.append(f"{site['site_id']}: no-candidate live segment did not use C1")
if b["labels"]:
if c["mode"] != "C2_candidate_conflict_adjudication":
errors.append(f"{site['site_id']}: candidate segment did not use C2")
if final == "no_training_like_candidate_detected_in_covered_live_segment":
if "negative_screen_coverage_sufficient" not in c["labels"]:
errors.append(f"{site['site_id']}: no-training final route lacks sufficient negative-screen coverage")
if "inconclusive_due_to_missingness" in c["labels"] and final == "no_training_like_candidate_detected_in_covered_live_segment":
errors.append(f"{site['site_id']}: missingness produced a false negative route")
if final in {"medium_training_like_warning", "high_training_like_warning"}:
if not ({"distributed_training_like_candidate", "checkpoint_training_like_candidate"} & set(b["labels"])):
errors.append(f"{site['site_id']}: warning height lacks identity-shape candidate")
return errors
def validate_website_data(website_dir: Path, results: dict[str, dict[str, Any]]) -> list[str]:
path = website_dir / "data" / "observable-demo-data.json"
if not path.exists():
return [f"website data missing: {path}"]
payload = json.loads(path.read_text(encoding="utf-8"))
errors = []
rows = payload.get("sites", [])
if len(rows) != len(results):
errors.append(f"website row count {len(rows)} does not match synthetic result count {len(results)}")
for row in rows:
site_id = row.get("site_id")
result = results.get(site_id)
if not result:
errors.append(f"website row {site_id!r} is not in synthetic sites")
continue
if row.get("result", {}).get("final_route") != result["final_route"]:
errors.append(f"website row {site_id}: final route does not match algorithm")
return errors
def validate_website_js(website_dir: Path) -> list[str]:
node = shutil.which("node")
if not node:
return ["node is not available for website JS validation"]
script = f"""
global.window = {{}};
require({json.dumps(str((website_dir / 'data' / 'observable-demo-data.js').resolve()))});
require({json.dumps(str((website_dir / 'scoring.js').resolve()))});
const data = window.DCVObservableDemoData;
const scoring = window.DCVObservableScoring;
const errors = [];
for (const row of data.sites) {{
const result = scoring.scoreState(row.demo_state);
if (result.final_route !== row.result.final_route) {{
errors.push(`${{row.site_id}} expected ${{row.result.final_route}} got ${{result.final_route}}`);
}}
}}
if (errors.length) {{
console.error(errors.join('\\n'));
process.exit(1);
}}
console.log(JSON.stringify({{rows: data.sites.length}}));
"""
proc = subprocess.run([node, "-e", script], text=True, capture_output=True, check=False)
if proc.returncode != 0:
return [f"website JS scoring mismatch: {proc.stderr.strip() or proc.stdout.strip()}"]
return []
if __name__ == "__main__":
main()
|