tools/check_standard.py

#!/usr/bin/env python3
import argparse
import re
from pathlib import Path

try:
    import yaml
except Exception as exc:
    raise SystemExit(f"PyYAML is required to parse YAML files: {exc}")


DATASET_ID = "OneScience/proteinmpnn"
MODEL_ID = "OneScience/ProteinMPNN"
REQUIRED_README_SECTIONS = [
    "## OneScience 官方信息",
    "## 项目说明",
    "## Resource Card",
    "## 文件说明",
    "## Manifest",
    "## 模型 vs 数据集关系",
    "## 文件与下载",
    "## 运行流程",
    "## 预检与诊断",
]
REQUIRED_MANIFEST_KEYS = [
    "resource",
    "platform_resource",
    "website_integration",
    "runtime",
    "onescience",
    "runtime_package",
    "files",
    "relations",
    "run_matrix",
    "capabilities",
    "commands",
    "expected_outputs",
    "diagnostics",
    "domain_extension",
]


def read_text_checked(path: Path) -> str:
    raw = path.read_bytes()
    try:
        text = raw.decode("utf-8")
    except UnicodeDecodeError as exc:
        raise SystemExit(f"{path} is not valid UTF-8: {exc}") from exc
    bad = []
    if "????" in text:
        bad.append("contains four consecutive question marks")
    if "\ufffd" in text:
        bad.append("contains U+FFFD replacement character")
    if re.search(r"(Ã|Â|Ð|Ñ|锟|鏂|璇|涓)", text):
        bad.append("contains possible mojibake markers")
    if bad:
        raise SystemExit(f"{path} encoding check failed: {', '.join(bad)}")
    return text


def cjk_count(text: str) -> int:
    return sum(1 for ch in text if "\u4e00" <= ch <= "\u9fff")


def walk_values(obj):
    if isinstance(obj, dict):
        for value in obj.values():
            yield from walk_values(value)
    elif isinstance(obj, list):
        for value in obj:
            yield from walk_values(value)
    else:
        yield obj


def collect_command_names(commands):
    names = set()
    if isinstance(commands, dict):
        for value in commands.values():
            if isinstance(value, list):
                for item in value:
                    if isinstance(item, dict) and "name" in item:
                        names.add(item["name"])
            elif isinstance(value, dict) and "name" in value:
                names.add(value["name"])
    return names


parser = argparse.ArgumentParser()
parser.add_argument("--root", default=".")
args = parser.parse_args()
root = Path(args.root).resolve()

readme = read_text_checked(root / "README.md")
manifest_text = read_text_checked(root / "manifest.yaml")
relations_text = read_text_checked(root / "onescience_relations.yaml")

if cjk_count(readme) < 200:
    raise SystemExit(f"README.md CJK count too low: {cjk_count(readme)}")
if cjk_count(manifest_text) < 20:
    raise SystemExit(f"manifest.yaml CJK count too low: {cjk_count(manifest_text)}")
for section in REQUIRED_README_SECTIONS:
    if section not in readme:
        raise SystemExit(f"README.md missing section: {section}")

manifest = yaml.safe_load(manifest_text)
relations = yaml.safe_load(relations_text)
missing = [k for k in REQUIRED_MANIFEST_KEYS if k not in manifest]
if missing:
    raise SystemExit(f"manifest missing keys: {missing}")

if manifest["resource"]["id"] != DATASET_ID:
    raise SystemExit("manifest resource.id mismatch")
primary = manifest["platform_resource"]["primary"]
if primary["repo_id"] != DATASET_ID or primary["repo_type"] != "dataset":
    raise SystemExit("manifest platform_resource.primary mismatch")

all_text = "\n".join([readme, manifest_text, relations_text])
if "Onescience/" in all_text or "onescience/" in all_text:
    raise SystemExit("found invalid OneScience namespace casing")
if " proteinmpnn" in all_text or "`proteinmpnn`" in all_text:
    pass
if "modelscope download --dataset proteinmpnn" in all_text:
    raise SystemExit("found bare dataset download id")
if "modelscope download --dataset OneScience/proteinmpnn" not in all_text:
    raise SystemExit("missing exact dataset download command")
if "OneScience/ProteinMPNN" not in all_text:
    raise SystemExit("missing compatible model id")

compatible = manifest["relations"].get("compatible_models", [])
if not compatible:
    raise SystemExit("missing relations.compatible_models")
for item in compatible:
    ref = item.get("resource_ref", {})
    if ref.get("repo_id") != MODEL_ID or ref.get("repo_type") != "model":
        raise SystemExit("compatible model resource_ref mismatch")

rel_dataset = relations.get("relations", {}).get("dataset", {}).get("resource_ref", {})
if rel_dataset.get("repo_id") != DATASET_ID:
    raise SystemExit("onescience_relations dataset repo_id mismatch")

command_names = collect_command_names(manifest.get("commands", {}))
if not command_names:
    raise SystemExit("no named commands found")
scenario_refs = [s.get("command_ref") for s in manifest.get("run_matrix", {}).get("scenarios", [])]
missing_refs = [ref for ref in scenario_refs if ref and ref not in command_names]
if missing_refs:
    raise SystemExit(f"run_matrix command_refs missing from commands: {missing_refs}")

for value in walk_values(manifest):
    if isinstance(value, str) and "repo_id" not in value:
        if "OneScience/proteinmpnn" in value or "OneScience/ProteinMPNN" in value:
            continue

print("STANDARD_CHECK_OK")
print(f"readme_cjk={cjk_count(readme)}")
print(f"manifest_cjk={cjk_count(manifest_text)}")
print(f"dataset_repo_id={DATASET_ID}")
print(f"compatible_model_repo_id={MODEL_ID}")
print("command_refs=PASS")