56cf2c3ee2
Landing the first province sample of the rules/_evidence/ tree.
Locks down the hunan-2026 本科批 院校专业组 rule set so audit
output and downstream consumers can cite the source摘录 instead
of just 'trust me, the loader says so'.
What's in the box:
- rules/_evidence/README.md: directory contract, evidence-file
template (4 required sections), coverage table, and a note on
RuleLoader's future strict_evidence toggle.
- rules/_evidence/hunan/<11 rule keys>.md: one Markdown
摘录 per source_evidence_id declared in hunan.yaml
(mode / batch / max_volunteers / max_majors_per_group /
has_adjustment / adjustment_scope / retrieval_rule /
collection_count / subject_mode / official_url /
exam_subject_total). Each file carries the official quote,
the machine-readable YAML form, key edge cases, and the
next-review date.
- tests/test_rules_evidence_layer.py: 4 invariants
* README exists and documents the contract
* every active hunan.yaml source_evidence_id has a matching
.md file
* 11 evidence files contain the keywords that prove they
actually carry the rule (parameterised, one test per file)
* every evidence file has all 4 required template sections
* no orphan evidence files (file exists but no rule
references it)
Verification:
- focused: 27 passed (15 evidence-layer + 12 cli doctor)
- dev-verify full gate: all checks passed (ruff / mypy /
coverage / pytest / benchmark)
Follow-up (not in this commit):
- Wire RuleLoader.from_truth_root(strict_evidence=True) so
callers can opt in to fail-loud on missing evidence
- Extend the same pattern to national.yaml + 2-3 more
provinces
- Add a 'last_verified_at' field to each rule and surface
>90-day-old evidence via 'gaokao-cli rules status'
105 lines
3.9 KiB
Python
105 lines
3.9 KiB
Python
from __future__ import annotations
|
|
|
|
from pathlib import Path
|
|
|
|
import pytest
|
|
|
|
PROJECT_ROOT = Path(__file__).resolve().parents[1]
|
|
|
|
|
|
def _evidence_root() -> Path:
|
|
return PROJECT_ROOT / "rules" / "_evidence"
|
|
|
|
|
|
def test_evidence_layer_readme_exists() -> None:
|
|
readme = _evidence_root() / "README.md"
|
|
assert readme.is_file(), "rules/_evidence/README.md must exist"
|
|
text = readme.read_text(encoding="utf-8")
|
|
# README must explain the directory contract.
|
|
assert "<prov-slug>" in text
|
|
assert "<source_evidence_id>" in text
|
|
assert "evidence 摘录模板" in text
|
|
|
|
|
|
def test_hunan_evidence_covers_every_active_rule() -> None:
|
|
"""Every source_evidence_id declared in rules/_truth/province/hunan.yaml
|
|
must have a matching evidence file under rules/_evidence/hunan/.
|
|
"""
|
|
import yaml # type: ignore[import-untyped]
|
|
|
|
truth_path = PROJECT_ROOT / "rules" / "_truth" / "province" / "hunan.yaml"
|
|
truth = yaml.safe_load(truth_path.read_text(encoding="utf-8")) or {}
|
|
rules = truth.get("rules", {}) or {}
|
|
expected_ids = {
|
|
payload["source_evidence_id"]
|
|
for payload in rules.values()
|
|
if payload.get("status", "active") == "active"
|
|
}
|
|
assert expected_ids, "hunan.yaml must declare at least one rule"
|
|
|
|
evidence_dir = _evidence_root() / "hunan"
|
|
missing: list[str] = []
|
|
for evidence_id in sorted(expected_ids):
|
|
if not (evidence_dir / f"{evidence_id}.md").is_file():
|
|
missing.append(evidence_id)
|
|
assert not missing, f"missing evidence files for: {missing}"
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"evidence_id,expected_keywords",
|
|
[
|
|
("hunan-2026-mode", ["院校专业组", "湖南省"]),
|
|
("hunan-2026-batch", ["本科批", "湖南省"]),
|
|
("hunan-2026-max_volunteers", ["45", "院校专业组"]),
|
|
("hunan-2026-max_majors_per_group", ["6", "专业"]),
|
|
("hunan-2026-has_adjustment", ["调剂", "院校专业组"]),
|
|
("hunan-2026-adjustment_scope", ["组内专业", "调剂"]),
|
|
("hunan-2026-retrieval_rule", ["分数优先", "一次投档"]),
|
|
("hunan-2026-collection_count", ["2", "征集"]),
|
|
("hunan-2026-subject_mode", ["3+1+2"]),
|
|
("hunan-2026-official_url", ["jyt.hunan.gov.cn"]),
|
|
("hunan-2026-exam_subject_total", ["750"]),
|
|
],
|
|
)
|
|
def test_hunan_evidence_files_contain_required_keywords(
|
|
evidence_id: str, expected_keywords: list[str]
|
|
) -> None:
|
|
path = _evidence_root() / "hunan" / f"{evidence_id}.md"
|
|
assert path.is_file(), f"missing evidence file: {path}"
|
|
text = path.read_text(encoding="utf-8")
|
|
for keyword in expected_keywords:
|
|
assert keyword in text, f"{evidence_id}.md must mention '{keyword}'"
|
|
|
|
|
|
def test_evidence_files_have_strict_template_sections() -> None:
|
|
"""Every evidence file must include the four required sections."""
|
|
for path in sorted((_evidence_root() / "hunan").glob("*.md")):
|
|
text = path.read_text(encoding="utf-8")
|
|
for section in (
|
|
"## 1. 官方原文摘录",
|
|
"## 2. 转写为机读规则",
|
|
"## 3. 关键边界与例外",
|
|
"## 4. 后续维护",
|
|
):
|
|
assert section in text, f"{path.name} missing section: {section}"
|
|
|
|
|
|
def test_no_orphan_evidence_files() -> None:
|
|
"""Every evidence file under rules/_evidence/hunan/ must be referenced
|
|
by some hunan.yaml rule. Prevents stale/orphan摘录.
|
|
"""
|
|
import yaml # type: ignore[import-untyped]
|
|
|
|
truth_path = PROJECT_ROOT / "rules" / "_truth" / "province" / "hunan.yaml"
|
|
truth = yaml.safe_load(truth_path.read_text(encoding="utf-8")) or {}
|
|
declared = {
|
|
payload["source_evidence_id"]
|
|
for payload in (truth.get("rules", {}) or {}).values()
|
|
if "source_evidence_id" in payload
|
|
}
|
|
evidence_files = {p.stem for p in (_evidence_root() / "hunan").glob("*.md")}
|
|
orphans = evidence_files - declared
|
|
assert not orphans, (
|
|
f"orphan evidence files (not referenced by hunan.yaml): {orphans}"
|
|
)
|