"""
Golden-set lint for masterclass community_search outputs.
Usage:
python tools/golden_eval.py path/to/output.md
cat output.md | python tools/golden_eval.py
The linter checks for required sections, evidence links with quotes,
and presence of code/commands. It is intentionally lightweight so it
can run in CI without network access.
"""
from __future__ import annotations
import json
import re
import sys
from pathlib import Path
from typing import Any, Dict, List
REQUIRED_SECTIONS = [
"## Findings",
"## Conflicts",
"## Recommended Path",
"## Quick-apply Code/Commands",
"## Verification",
"## Search Stats",
]
def lint_masterclass_output(text: str) -> Dict[str, Any]:
"""Return lint findings for a masterclass-style markdown output."""
issues: List[str] = []
metrics: Dict[str, Any] = {}
for section in REQUIRED_SECTIONS:
if section not in text:
issues.append(f"Missing section: {section}")
evidence_links = re.findall(r"https?://\\S+", text)
if len(evidence_links) < 2:
issues.append("Evidence weak: fewer than 2 links detected.")
metrics["evidence_links"] = len(evidence_links)
quotes = re.findall(r"\"[^\"]{10,200}\"", text)
if not quotes:
issues.append("No quoted evidence snippets found.")
metrics["quotes"] = len(quotes)
code_blocks = text.count("```")
if code_blocks < 2:
issues.append("No code/command block detected.")
metrics["code_blocks"] = code_blocks // 2
findings_section = re.search(r"## Findings[\\s\\S]+?##", text)
if findings_section:
findings_count = len(re.findall(r"\\n\\d+\\)", findings_section.group(0)))
metrics["findings_count"] = findings_count
if findings_count == 0:
issues.append("No enumerated findings detected.")
return {
"ok": len(issues) == 0,
"issues": issues,
"metrics": metrics,
}
def main() -> int:
if len(sys.argv) > 1:
content = Path(sys.argv[1]).read_text(encoding="utf-8")
else:
content = sys.stdin.read()
report = lint_masterclass_output(content)
print(json.dumps(report, indent=2))
return 0 if report["ok"] else 1
if __name__ == "__main__":
sys.exit(main())
