code/scripts/grl_license_check.py

#!/usr/bin/env python3
"""Check CC BY 4.0 license metadata for the GRL submission package."""

from __future__ import annotations

import json
import re
from pathlib import Path


ROOT = Path(__file__).resolve().parents[1]
PUBLISH_ROOT = ROOT.parent / "grl_publish"
OUT = ROOT / "submission_final"


def read(path: Path) -> str:
    return path.read_text(encoding="utf-8", errors="replace") if path.exists() else ""


def has(path: Path, needle: str) -> bool:
    return needle.lower() in read(path).lower()


def yaml_license_ok(readme: Path) -> bool:
    text = read(readme)
    if not text.startswith("---"):
        return False
    parts = text.split("---", 2)
    if len(parts) < 3:
        return False
    return re.search(r"(?mi)^license:\s*cc-by-4\.0\s*$", parts[1]) is not None


def main() -> int:
    OUT.mkdir(parents=True, exist_ok=True)
    failures: list[str] = []
    warnings: list[str] = []

    main_tex = ROOT / "main.tex"
    readme_reproduce = ROOT / "README_reproduce.md"
    citation = ROOT / "CITATION.cff"
    license_file = ROOT / "LICENSE"
    notice_file = ROOT / "NOTICE"
    publish_readme = PUBLISH_ROOT / "README.md"
    publish_license = PUBLISH_ROOT / "LICENSE"
    publish_notice = PUBLISH_ROOT / "NOTICE"
    publish_citation = PUBLISH_ROOT / "CITATION.cff"

    for path in [license_file, publish_license]:
        if not path.exists():
            failures.append(f"Missing LICENSE: {path}")
        elif not has(path, "CC-BY-4.0") or not has(path, "https://creativecommons.org/licenses/by/4.0/"):
            failures.append(f"LICENSE does not contain CC-BY-4.0 and license URL: {path}")

    for path in [notice_file, publish_notice]:
        if not path.exists():
            failures.append(f"Missing NOTICE: {path}")
        elif not has(path, "snr_bias") or not has(path, "CC BY 4.0"):
            failures.append(f"NOTICE missing snr_bias archive URL/name or CC BY 4.0: {path}")

    if not has(readme_reproduce, "CC BY 4.0"):
        failures.append("README_reproduce.md does not contain CC BY 4.0.")
    if not has(publish_readme, "CC BY 4.0"):
        failures.append("grl_publish/README.md does not contain CC BY 4.0.")
    if not yaml_license_ok(publish_readme):
        failures.append("grl_publish/README.md Hugging Face YAML metadata missing license: cc-by-4.0.")

    for path in [citation, publish_citation]:
        if not path.exists():
            failures.append(f"Missing CITATION.cff: {path}")
        elif not has(path, "CC-BY-4.0") or not has(path, "snr_bias"):
            failures.append(f"CITATION.cff missing CC-BY-4.0 or snr_bias archive URL/name: {path}")

    manuscript = read(main_tex)
    forbidden = [
        "no separate license file",
        "license file was present",
        "TBD",
        "placeholder",
        "Requires author confirmation",
        "email to be added",
        "Affiliation to be added",
    ]
    for token in forbidden:
        if token.lower() in manuscript.lower():
            failures.append(f"Forbidden manuscript text remains: {token}")

    required_open_research = [
        "CC BY 4.0",
        "snr_bias",
        "Hugging Face",
        "SeismicX-Cont",
        "DOI: 10.57967/hf/9006",
        "SeisDispFusion-NCF",
        "DOI: 10.57967/hf/9114",
        r"\citeA{yu2026snrbias}",
        r"\citeA{yu2026seismicxcont}",
        r"\citeA{yu2026seisdispfusionncf}",
    ]
    for token in required_open_research:
        if token not in manuscript:
            failures.append(f"Open Research missing required text: {token}")

    bad_creditx1 = [
        "CREDIT-X1local data are redistributed",
        "CREDIT-X1local waveforms are redistributed",
        "raw CREDIT-X1local data are included",
    ]
    for token in bad_creditx1:
        if token.lower() in manuscript.lower():
            failures.append(f"Potential false redistribution claim: {token}")

    if not PUBLISH_ROOT.exists():
        warnings.append(f"Publish archive root not found: {PUBLISH_ROOT}")

    status = "PASS" if not failures else "FAIL"
    report = {
        "overall_status": status,
        "failures": failures,
        "warnings": warnings,
        "checked_paths": {
            "manuscript_root": str(ROOT),
            "publish_root": str(PUBLISH_ROOT),
            "main_tex": str(main_tex),
            "publish_readme": str(publish_readme),
        },
    }
    (OUT / "license_check.json").write_text(json.dumps(report, indent=2) + "\n", encoding="utf-8")

    lines = [
        "# License Check",
        "",
        f"- Overall status: `{status}`",
        "",
        "## Failures",
        "",
        *([f"- {item}" for item in failures] or ["- None"]),
        "",
        "## Warnings",
        "",
        *([f"- {item}" for item in warnings] or ["- None"]),
    ]
    (OUT / "license_check.md").write_text("\n".join(lines) + "\n", encoding="utf-8")
    print(json.dumps(report, indent=2))
    return 0 if status == "PASS" else 1


if __name__ == "__main__":
    raise SystemExit(main())