#!/usr/bin/env python3
"""Fail when source files exceed a configured line-count threshold."""

from __future__ import annotations

import argparse
from pathlib import Path


DEFAULT_SKIP_PARTS = {
    ".git",
    ".venv",
    "venv",
    "build",
    "dist",
    "node_modules",
    "__pycache__",
    ".pytest_cache",
}
SOURCE_SUFFIXES = {".py", ".sh", ".json", ".yaml", ".yml"}


def _read_waivers(path: Path) -> set[str]:
    if not path.exists():
        return set()
    waived: set[str] = set()
    for line in path.read_text(encoding="utf-8").splitlines():
        row = line.strip()
        if not row or row.startswith("#"):
            continue
        waived.add(row.split("\t", 1)[0].strip())
    return waived


def _iter_files(root: Path) -> list[Path]:
    if not root.exists():
        return []
    files: list[Path] = []
    for path in root.rglob("*"):
        if not path.is_file():
            continue
        if any(part in DEFAULT_SKIP_PARTS for part in path.parts):
            continue
        if path.suffix.lower() not in SOURCE_SUFFIXES and path.name != "Jenkinsfile":
            continue
        files.append(path)
    return files


def main() -> int:
    parser = argparse.ArgumentParser(description=__doc__)
    parser.add_argument("--roots", nargs="+", required=True)
    parser.add_argument("--max-lines", type=int, default=500)
    parser.add_argument("--waivers", default="ci/loc_hygiene_waivers.tsv")
    args = parser.parse_args()

    repo_root = Path.cwd().resolve()
    waived = _read_waivers(repo_root / args.waivers)

    offenders: list[tuple[int, str]] = []
    for root_name in args.roots:
        for path in _iter_files(repo_root / root_name):
            rel = path.relative_to(repo_root).as_posix()
            if rel in waived:
                continue
            try:
                line_count = sum(1 for _ in path.open("r", encoding="utf-8", errors="ignore"))
            except OSError:
                continue
            if line_count > args.max_lines:
                offenders.append((line_count, rel))

    if not offenders:
        print(f"[loc] ok: no files exceed {args.max_lines} lines")
        return 0

    offenders.sort(reverse=True)
    print(f"[loc] failed: {len(offenders)} file(s) exceed {args.max_lines} lines")
    for lines, rel in offenders:
        print(f" - {rel}: {lines} lines")
    return 1


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