fix: preserve world-class submission directory context

This commit is contained in:
yaojingang
2026-06-14 12:08:54 +08:00
parent 22c7145db7
commit e03d5b2c74
6 changed files with 33 additions and 6 deletions
@@ -2,6 +2,7 @@
import argparse
import html
import json
import shlex
import shutil
from datetime import date
from pathlib import Path
@@ -22,6 +23,10 @@ def rel_path(path: Path, root: Path) -> str:
return str(path.resolve())
def shell_path(path: Path, root: Path) -> str:
return shlex.quote(rel_path(path, root))
def write_json(path: Path, payload: dict[str, Any]) -> None:
path.parent.mkdir(parents=True, exist_ok=True)
path.write_text(json.dumps(payload, ensure_ascii=False, indent=2) + "\n", encoding="utf-8")
@@ -338,8 +343,8 @@ def build_submission_kit(
"files": files,
"evidence_items": items,
"commands": {
"validate_intake": f"python3 scripts/yao.py world-class-intake . --submissions-dir {rel_path(output_dir, skill_dir)}",
"refresh_ledger": f"python3 scripts/yao.py world-class-ledger . --submissions-dir {rel_path(output_dir, skill_dir)}",
"validate_intake": f"python3 scripts/yao.py world-class-intake . --submissions-dir {shell_path(output_dir, skill_dir)}",
"refresh_ledger": f"python3 scripts/yao.py world-class-ledger . --submissions-dir {shell_path(output_dir, skill_dir)}",
"guard_claim": "python3 scripts/yao.py world-class-claim-guard .",
},
"safety": {
@@ -101,7 +101,10 @@ def build_operator_checklist(
f"--evidence-key {shlex.quote(key)} --output-dir {shell_path(submission_path.parent, skill_dir)}"
),
"validate_intake": f"python3 scripts/yao.py world-class-intake . --submissions-dir {submissions_dir_arg}",
"refresh_ledger": "python3 scripts/yao.py world-class-ledger .",
"submission_review": (
f"python3 scripts/yao.py world-class-submission-review . --submissions-dir {submissions_dir_arg}"
),
"refresh_ledger": f"python3 scripts/yao.py world-class-ledger . --submissions-dir {submissions_dir_arg}",
"guard_claim": "python3 scripts/yao.py world-class-claim-guard .",
},
"must_collect": {
@@ -255,7 +258,7 @@ def render_operator_checklist(items: list[dict[str, Any]]) -> list[str]:
lines.append(f"- submission: `{item['submission_path']}`")
lines.extend(["", "#### Commands", ""])
commands = item.get("commands", {})
for label in ["prepare_submission", "validate_intake", "refresh_ledger", "guard_claim"]:
for label in ["prepare_submission", "validate_intake", "submission_review", "refresh_ledger", "guard_claim"]:
if commands.get(label):
lines.append(f"- {label}: `{commands[label]}`")
must_collect = item.get("must_collect", {})
@@ -62,7 +62,7 @@ def build_runbook_item(
"commands": {
"prepare_submission": commands.get("prepare_submission", ""),
"validate_intake": commands.get("validate_intake", ""),
"review_queue": "python3 scripts/yao.py world-class-submission-review . --submissions-dir evidence/world_class/submissions",
"review_queue": commands.get("submission_review", ""),
"refresh_ledger": commands.get("refresh_ledger", "python3 scripts/yao.py world-class-ledger ."),
"guard_claim": commands.get("guard_claim", "python3 scripts/yao.py world-class-claim-guard ."),
},
+2
View File
@@ -493,6 +493,8 @@ def main() -> None:
assert provider_checklist["readiness"] == "awaiting-submission", provider_checklist
assert provider_checklist["submission_path"] == "evidence/world_class/submissions/provider-holdout.json", provider_checklist
assert provider_checklist["commands"]["validate_intake"] == "python3 scripts/yao.py world-class-intake . --submissions-dir evidence/world_class/submissions", provider_checklist
assert provider_checklist["commands"]["submission_review"] == "python3 scripts/yao.py world-class-submission-review . --submissions-dir evidence/world_class/submissions", provider_checklist
assert provider_checklist["commands"]["refresh_ledger"] == "python3 scripts/yao.py world-class-ledger . --submissions-dir evidence/world_class/submissions", provider_checklist
assert "provider-backed model run" in provider_checklist["must_collect"]["provenance_requirements"], provider_checklist
assert full_payload["data"]["world_class_claim_guard"]["summary"]["decision"] == "claim-guard-pass-evidence-pending", full_payload["data"]["world_class_claim_guard"]
assert full_payload["data"]["world_class_claim_guard"]["summary"]["violation_count"] == 0, full_payload["data"]["world_class_claim_guard"]
+15 -1
View File
@@ -146,6 +146,8 @@ def main() -> None:
"--evidence-key provider-holdout --output-dir evidence/world_class/submissions"
), checklist["provider-holdout"]
assert checklist["provider-holdout"]["commands"]["validate_intake"] == "python3 scripts/yao.py world-class-intake . --submissions-dir evidence/world_class/submissions", checklist["provider-holdout"]
assert checklist["provider-holdout"]["commands"]["submission_review"] == "python3 scripts/yao.py world-class-submission-review . --submissions-dir evidence/world_class/submissions", checklist["provider-holdout"]
assert checklist["provider-holdout"]["commands"]["refresh_ledger"] == "python3 scripts/yao.py world-class-ledger . --submissions-dir evidence/world_class/submissions", checklist["provider-holdout"]
assert "provider-backed model run" in checklist["provider-holdout"]["must_collect"]["provenance_requirements"], checklist["provider-holdout"]
assert "reports/output_execution_runs.json summary.model_executed_count > 0" in checklist["provider-holdout"]["must_collect"]["success_checks"], checklist["provider-holdout"]
assert checklist["provider-holdout"]["anti_overclaim"]["local_command_runner_counts_as_provider_model"] is False, checklist["provider-holdout"]
@@ -158,7 +160,8 @@ def main() -> None:
assert "0 existing / 0 sha256 verified / 0 required verified / 1 refs" in markdown, markdown
assert "`evidence/world_class/submissions/provider-holdout.json`" in markdown, markdown
assert "`python3 scripts/yao.py world-class-submission-kit . --evidence-key provider-holdout --output-dir evidence/world_class/submissions`" in markdown, markdown
assert "`python3 scripts/yao.py world-class-ledger .`" in markdown, markdown
assert "`python3 scripts/yao.py world-class-submission-review . --submissions-dir evidence/world_class/submissions`" in markdown, markdown
assert "`python3 scripts/yao.py world-class-ledger . --submissions-dir evidence/world_class/submissions`" in markdown, markdown
assert "Templates and planned work do not count as accepted evidence." in markdown, markdown
assert "Real submissions must include the evidence-key critical artifact paths with verified SHA-256 digests." in markdown, markdown
assert "Real submissions must replace template submitter, date, and provenance placeholders with concrete evidence metadata." in markdown, markdown
@@ -196,6 +199,15 @@ def main() -> None:
assert draft_intake["submissions"][0]["evidence_key"] == "provider-holdout", draft_intake["submissions"]
assert all(item["evidence_key"] != "unknown" for item in draft_intake["submissions"]), draft_intake["submissions"]
spaced_kit_dir = TMP / "submission kit spaced"
spaced_kit_proc = run_kit("--output-dir", str(spaced_kit_dir), "--evidence-key", "provider-holdout")
spaced_kit_payload = json.loads(spaced_kit_proc.stdout)
quoted_spaced_dir = "'tests/tmp_world_class_evidence_intake/submission kit spaced'"
assert quoted_spaced_dir in spaced_kit_payload["commands"]["validate_intake"], spaced_kit_payload["commands"]
assert quoted_spaced_dir in spaced_kit_payload["commands"]["refresh_ledger"], spaced_kit_payload["commands"]
spaced_readme = (spaced_kit_dir / "README.md").read_text(encoding="utf-8")
assert quoted_spaced_dir in spaced_readme, spaced_readme
existing_proc = run_kit("--output-dir", str(kit_dir), "--evidence-key", "provider-holdout")
existing_payload = json.loads(existing_proc.stdout)
assert existing_payload["summary"]["existing_count"] == 1, existing_payload["summary"]
@@ -223,6 +235,8 @@ def main() -> None:
assert "source evidence checks still do not pass" in valid_checklist["provider-holdout"]["blocking_reason"], valid_checklist["provider-holdout"]
assert valid_checklist["provider-holdout"]["submission_path"].endswith("tests/tmp_world_class_evidence_intake/valid_submissions/provider-holdout.json"), valid_checklist["provider-holdout"]
assert "tests/tmp_world_class_evidence_intake/valid_submissions" in valid_checklist["provider-holdout"]["commands"]["validate_intake"], valid_checklist["provider-holdout"]
assert "tests/tmp_world_class_evidence_intake/valid_submissions" in valid_checklist["provider-holdout"]["commands"]["submission_review"], valid_checklist["provider-holdout"]
assert "tests/tmp_world_class_evidence_intake/valid_submissions" in valid_checklist["provider-holdout"]["commands"]["refresh_ledger"], valid_checklist["provider-holdout"]
placeholder_dir = TMP / "placeholder_submissions"
placeholder_dir.mkdir()
@@ -167,6 +167,9 @@ def main() -> None:
assert submitted_provider["intake_readiness"] == "source-evidence-incomplete", submitted_provider
assert submitted_provider["review_state"] == "source-evidence-incomplete", submitted_provider
assert submitted_provider["source_accepted"] is False, submitted_provider
assert "tests/tmp_world_class_operator_runbook/valid_submissions" in submitted_provider["commands"]["validate_intake"], submitted_provider
assert "tests/tmp_world_class_operator_runbook/valid_submissions" in submitted_provider["commands"]["review_queue"], submitted_provider
assert "tests/tmp_world_class_operator_runbook/valid_submissions" in submitted_provider["commands"]["refresh_ledger"], submitted_provider
print(json.dumps({"ok": True}, ensure_ascii=False, indent=2))