check_comment_to_code_ratio.py

import argparse
import os
import sys
from pathlib import Path

EXPECTED_HEADER_SUFFIX = [
    "== files",
    "!= files",
    "+ files",
    "- files",
    "== blank",
    "!= blank",
    "+ blank",
    "- blank",
    "== comment",
    "!= comment",
    "+ comment",
    "- comment",
    "== code",
    "!= code",
    "+ code",
    "- code",
]
# Zero-based indices in `cloc --diff --md` rows after splitting by `|`.
COMMENT_COLUMN_INDICES = (10, 11, 12)
CODE_COLUMN_INDICES = (14, 15, 16)


def parse_changed_totals(report_path):
    changed_comments = 0
    changed_code = 0
    header_found = False

    for raw_line in report_path.read_text(encoding="utf-8").splitlines():
        if "|" not in raw_line:
            continue

        cells = [cell.strip() for cell in raw_line.split("|")]
        if cells and cells[-1] == "":
            cells = cells[:-1]

        if cells and cells[0] in {"Language", "File"}:
            if cells[1:] != EXPECTED_HEADER_SUFFIX:
                sys.exit("Unexpected cloc diff table format.")
            header_found = True
            continue

        if not cells:
            continue
        if cells[0].startswith((":", "-", "SUM:")):
            continue
        if not header_found:
            continue

        try:
            changed_comments += sum(
                int(cells[index]) for index in COMMENT_COLUMN_INDICES
            )
            changed_code += sum(int(cells[index]) for index in CODE_COLUMN_INDICES)
        except (IndexError, ValueError):
            continue

    if not header_found:
        print("Unable to find cloc diff table header in report.")
        sys.exit(1)

    return changed_comments, changed_code


def main():
    parser = argparse.ArgumentParser()
    parser.add_argument("report")
    parser.add_argument("--limit", type=float, default=10.0)
    args = parser.parse_args()

    report_path = Path(args.report)
    if not report_path.exists():
        print(f"{report_path} was not generated by the comment ratio action.")
        sys.exit(1)

    changed_comments, changed_code = parse_changed_totals(report_path)
    total_changed_nonblank = changed_comments + changed_code
    comment_percentage = (
        0.0
        if total_changed_nonblank == 0
        else changed_comments * 100.0 / total_changed_nonblank
    )

    print(f"Comment percentage for PR changes: {comment_percentage:.2f}%")

    github_output = os.getenv("GITHUB_OUTPUT")
    if github_output:
        with Path(github_output).open("a", encoding="utf-8") as output:
            output.write(f"changed_comments={changed_comments}\n")
            output.write(f"changed_code={changed_code}\n")
            output.write(f"comment_percentage={comment_percentage:.2f}\n")

    if comment_percentage > args.limit:
        print(
            f"Comment percentage for PR changes is {comment_percentage:.2f}%, "
            f"which exceeds {args.limit:.2f}%.",
            file=sys.stderr,
        )
        sys.exit(1)


if __name__ == "__main__":
    main()