Datasets:
Download scripts/check_docs.py from yamahei21/uzu-quant-kit: direct link, hf CLI and curl.
- Browser
- Download file 4.96 kB
-
https://huggingface.co/datasets/yamahei21/uzu-quant-kit/resolve/main/scripts/check_docs.py
- Command line
-
hf download hf://datasets/yamahei21/uzu-quant-kit/scripts/check_docs.py
-
curl -L -o check_docs.py https://huggingface.co/datasets/yamahei21/uzu-quant-kit/resolve/main/scripts/check_docs.py
4.96 kB
| #!/usr/bin/env python3 | |
| """Check that commands in the docs still match the scripts. | |
| Docs drift caught us twice: RUNBOOK E2/E3 lacked --eval-corpus (the guard exits | |
| immediately), and Step 3 lacked --calibration-json (provenance becomes | |
| overlap_checked=false and QAD later refuses the model). Both were executable | |
| lines in markdown. | |
| This scans fenced code blocks for ``python scripts/<name>.py ...`` lines, then | |
| verifies: | |
| - the script exists | |
| - every ``--flag`` is declared in that script's ``add_argument`` calls | |
| - ``--spec lanes/*.json`` points at an existing file | |
| Usage: | |
| python scripts/check_docs.py | |
| python scripts/check_docs.py --strict | |
| python scripts/check_docs.py --root . --file experiments/qwen35-0.8b/RUNBOOK.md | |
| """ | |
| from __future__ import annotations | |
| import argparse | |
| import re | |
| from pathlib import Path | |
| ROOT = Path(__file__).resolve().parents[1] | |
| DEFAULT_DOCS = [ | |
| "README.md", | |
| "experiments/PROTOCOL.md", | |
| "experiments/qwen35-0.8b/RUNBOOK.md", | |
| "experiments/qwen35-0.8b/EXPERIMENT_MATRIX.md", | |
| "lanes/README.md", | |
| "docs/LOCAL_POLICY.md", | |
| "docs/KAGGLE_SLOTS.md", | |
| "docs/EVAL.md", | |
| "docs/MODAL.md", | |
| "docs/ENVIRONMENT.md", | |
| "docs/STATUS.md", | |
| "tests/README.md", | |
| "antigravity/README.md", | |
| ] | |
| FENCE = re.compile(r"```(?:bash|sh)\n(.*?)```", re.S) | |
| COMMAND = re.compile(r"(?:^|\s)(?:[A-Z_]+=\S+\s+)*python3?\s+(scripts/[\w.-]+\.py)([^\n|]*)") | |
| def code_blocks(text: str) -> list[tuple[int, str]]: | |
| blocks = [] | |
| for match in FENCE.finditer(text): | |
| line = text[: match.start()].count("\n") + 2 | |
| blocks.append((line, match.group(1))) | |
| return blocks | |
| def extract_commands(text: str) -> list[dict]: | |
| commands = [] | |
| for block_line, block in code_blocks(text): | |
| for offset, line in enumerate(block.splitlines()): | |
| stripped = line.strip() | |
| if not stripped or stripped.startswith("#"): | |
| continue | |
| for match in COMMAND.finditer(stripped): | |
| commands.append( | |
| { | |
| "line": block_line + offset, | |
| "script": match.group(1), | |
| "args": match.group(2).strip(), | |
| } | |
| ) | |
| return commands | |
| def declaration_flags(script: Path) -> set[str]: | |
| text = script.read_text(encoding="utf-8") | |
| return set(re.findall(r'["\'](--[a-z0-9][a-z0-9-]*)["\']', text)) | |
| def command_flags(args: str) -> set[str]: | |
| # stop at a shell pipe/redirect | |
| cleaned = re.split(r"[|>]", args)[0] | |
| return set(re.findall(r"(--[a-z0-9][a-z0-9-]*)", cleaned)) | |
| def spec_paths(args: str) -> list[str]: | |
| cleaned = re.split(r"[|>]", args)[0] | |
| return re.findall(r"--spec\s+(\S+)", cleaned) | |
| def check(root: Path, files: list[str]) -> list[dict]: | |
| problems: list[dict] = [] | |
| for name in files: | |
| doc = root / name | |
| if not doc.exists(): | |
| continue | |
| for command in extract_commands(doc.read_text(encoding="utf-8")): | |
| script = root / command["script"] | |
| if not script.exists(): | |
| problems.append( | |
| {"doc": name, "line": command["line"], "kind": "missing-script", "detail": command["script"]} | |
| ) | |
| continue | |
| known = declaration_flags(script) | |
| unknown = sorted(command_flags(command["args"]) - known) | |
| if unknown: | |
| problems.append( | |
| { | |
| "doc": name, | |
| "line": command["line"], | |
| "kind": "unknown-flag", | |
| "detail": f"{command['script']}: {', '.join(unknown)}", | |
| } | |
| ) | |
| for spec in spec_paths(command["args"]): | |
| if spec.startswith("$") or spec.startswith("/"): | |
| continue | |
| if not (root / spec).exists(): | |
| problems.append( | |
| {"doc": name, "line": command["line"], "kind": "missing-spec", "detail": spec} | |
| ) | |
| return problems | |
| def main() -> None: | |
| parser = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter) | |
| parser.add_argument("--root", type=Path, default=ROOT) | |
| parser.add_argument("--file", action="append", default=None, help="check only these docs") | |
| parser.add_argument("--strict", action="store_true") | |
| args = parser.parse_args() | |
| files = args.file or DEFAULT_DOCS | |
| problems = check(args.root, files) | |
| checked = sum(1 for name in files if (args.root / name).exists()) | |
| print(f"checked {checked} docs") | |
| if not problems: | |
| print("DOCS OK") | |
| return | |
| for problem in problems: | |
| print(f"{problem['doc']}:{problem['line']} [{problem['kind']}] {problem['detail']}") | |
| print(f"{len(problems)} problem(s)") | |
| if args.strict: | |
| raise SystemExit(1) | |
| if __name__ == "__main__": | |
| main() | |