#!/usr/bin/env python3
"""本機 repo canonical lineage 只讀探測。

此工具比較多個本機 Git working tree 的 HEAD、branch、remote 與近期
commit ancestry，協助判斷它們是否可能屬於同一個 canonical repo。
它不 fetch、不push、不修改 remote，也不讀取 commit message。
"""

from __future__ import annotations

import argparse
import configparser
import json
import subprocess
from pathlib import Path
from urllib.parse import urlsplit, urlunsplit


def redact_url(value: str) -> str:
    if "://" not in value:
        if "@" in value and ":" in value.split("@", 1)[1]:
            return value.split("@", 1)[1]
        return value
    parts = urlsplit(value)
    netloc = parts.netloc.split("@", 1)[-1]
    return urlunsplit((parts.scheme, netloc, parts.path, parts.query, parts.fragment))


def run_git(repo: Path, args: list[str], timeout: int) -> subprocess.CompletedProcess[str]:
    try:
        return subprocess.run(
            ["git", *args],
            cwd=repo,
            check=False,
            capture_output=True,
            text=True,
            timeout=timeout,
        )
    except subprocess.TimeoutExpired:
        return subprocess.CompletedProcess(["git", *args], 124, "", "git command timeout")


def git_value(repo: Path, args: list[str], timeout: int) -> str:
    result = run_git(repo, args, timeout)
    if result.returncode != 0:
        return ""
    return result.stdout.strip()


def git_config_path(repo_path: Path) -> Path | None:
    git_path = repo_path / ".git"
    if git_path.is_dir():
        config_path = git_path / "config"
        return config_path if config_path.exists() else None
    if not git_path.is_file():
        return None
    text = git_path.read_text(encoding="utf-8", errors="replace")
    for line in text.splitlines():
        if line.startswith("gitdir:"):
            raw_gitdir = line.split(":", 1)[1].strip()
            gitdir = Path(raw_gitdir)
            if not gitdir.is_absolute():
                gitdir = (repo_path / gitdir).resolve()
            config_path = gitdir / "config"
            return config_path if config_path.exists() else None
    return None


def remote_name(section: str) -> str | None:
    prefix = 'remote "'
    if section.startswith(prefix) and section.endswith('"'):
        return section[len(prefix) : -1]
    return None


def read_remotes(repo_path: Path) -> list[dict[str, str]]:
    config_path = git_config_path(repo_path)
    if config_path is None:
        return []
    parser = configparser.RawConfigParser(strict=False)
    parser.read(config_path, encoding="utf-8")
    remotes: list[dict[str, str]] = []
    for section in parser.sections():
        name = remote_name(section)
        if not name or not parser.has_option(section, "url"):
            continue
        remotes.append(
            {
                "name": name,
                "url_redacted": redact_url(parser.get(section, "url").strip()),
            }
        )
    return remotes


def parse_repo_arg(value: str) -> tuple[str, Path]:
    if "=" not in value:
        raise argparse.ArgumentTypeError("--repo 必須是 label=/absolute/path")
    label, raw_path = value.split("=", 1)
    if not label.strip() or not raw_path.strip():
        raise argparse.ArgumentTypeError("--repo label 與 path 不可為空")
    return label.strip(), Path(raw_path).expanduser().resolve()


def repo_summary(label: str, repo_path: Path, sample_limit: int, git_timeout: int) -> dict[str, object]:
    exists = (repo_path / ".git").exists()
    if not exists:
        return {
            "label": label,
            "repo_path": str(repo_path),
            "exists": False,
            "head_sha": "",
            "head_short": "",
            "branch": "",
            "commit_sample_count": 0,
            "commits": [],
            "remotes": [],
            "probe_error": "repo missing",
        }

    probe_errors: list[str] = []
    head_result = run_git(repo_path, ["rev-parse", "HEAD"], git_timeout)
    head_sha = head_result.stdout.strip() if head_result.returncode == 0 else ""
    if head_result.returncode != 0:
        probe_errors.append("HEAD 讀取失敗或逾時")

    branch = git_value(repo_path, ["rev-parse", "--abbrev-ref", "HEAD"], git_timeout)
    rev_list_result = run_git(repo_path, ["rev-list", f"--max-count={sample_limit}", "HEAD"], git_timeout)
    commits = rev_list_result.stdout.splitlines() if rev_list_result.returncode == 0 else []
    if rev_list_result.returncode != 0:
        probe_errors.append("rev-list 讀取失敗或逾時")
    return {
        "label": label,
        "repo_path": str(repo_path),
        "exists": True,
        "head_sha": head_sha,
        "head_short": head_sha[:7],
        "branch": branch,
        "commit_sample_count": len(commits),
        "commits": commits,
        "remotes": read_remotes(repo_path),
        "probe_error": "；".join(probe_errors),
    }


def compare_repos(left: dict[str, object], right: dict[str, object]) -> dict[str, object]:
    left_commits = set(left.get("commits", []))
    right_commits = set(right.get("commits", []))
    left_head = str(left.get("head_sha") or "")
    right_head = str(right.get("head_sha") or "")
    common = sorted(left_commits & right_commits)

    if not left.get("exists") or not right.get("exists"):
        relation = "missing_repo"
    elif left_head and left_head == right_head:
        relation = "same_head"
    elif left.get("probe_error") or right.get("probe_error"):
        relation = "partial_probe"
    elif right_head and right_head in left_commits:
        relation = "left_descends_from_right"
    elif left_head and left_head in right_commits:
        relation = "right_descends_from_left"
    elif common:
        relation = "shared_history"
    else:
        relation = "no_shared_history"

    return {
        "left_label": left["label"],
        "right_label": right["label"],
        "relation": relation,
        "left_head": left_head,
        "right_head": right_head,
        "common_commit_count": len(common),
        "common_commit_samples": common[:5],
    }


def build_payload(
    group_name: str,
    repo_args: list[tuple[str, Path]],
    sample_limit: int,
    git_timeout: int,
) -> dict[str, object]:
    repos = [repo_summary(label, path, sample_limit, git_timeout) for label, path in repo_args]
    comparisons = []
    for left_index, left in enumerate(repos):
        for right in repos[left_index + 1 :]:
            comparisons.append(compare_repos(left, right))
    partial = any(item["relation"] == "partial_probe" for item in comparisons)
    related = any(
        item["relation"]
        in ("same_head", "left_descends_from_right", "right_descends_from_left", "shared_history")
        for item in comparisons
    )
    no_shared = any(item["relation"] == "no_shared_history" for item in comparisons)
    if partial:
        status = "partial"
    elif related and no_shared:
        status = "mixed"
    elif related:
        status = "related"
    elif comparisons:
        status = "unrelated"
    else:
        status = "partial"
    return {
        "schema_version": "local_repo_canonical_probe_v1",
        "group_name": group_name,
        "status": status,
        "sample_limit": sample_limit,
        "git_timeout_seconds": git_timeout,
        "repo_count": len(repos),
        "comparison_count": len(comparisons),
        "repos": repos,
        "comparisons": comparisons,
    }


def write_markdown(payload: dict[str, object], path: Path) -> None:
    lines = [
        "# 本機 Repo Canonical Lineage Probe 快照",
        "",
        "| 項目 | 值 |",
        "|------|----|",
        f"| 群組 | `{payload['group_name']}` |",
        f"| 狀態 | `{payload['status']}` |",
        f"| repo 數 | `{payload['repo_count']}` |",
        f"| 比對數 | `{payload['comparison_count']}` |",
        f"| sample limit | `{payload['sample_limit']}` |",
        f"| git timeout seconds | `{payload['git_timeout_seconds']}` |",
        "",
        "## Repo HEAD",
        "",
        "| Label | Path | Branch | HEAD | Remotes |",
        "|-------|------|--------|------|---------|",
    ]
    for repo in payload.get("repos", []):
        if not isinstance(repo, dict):
            continue
        remotes = repo.get("remotes", [])
        remote_text = ", ".join(
            f"`{remote.get('name', '')}:{remote.get('url_redacted', '')}`"
            for remote in remotes
            if isinstance(remote, dict)
        )
        lines.append(
            "| "
            + " | ".join(
                [
                    f"`{repo.get('label', '')}`",
                    f"`{repo.get('repo_path', '')}`",
                    f"`{repo.get('branch', '')}`",
                    f"`{repo.get('head_short', '')}`",
                    remote_text or "-",
                ]
            )
            + " |"
        )

    lines.extend(["", "## Lineage 比對", "", "| Left | Right | Relation | Common commits |", "|------|-------|----------|----------------|"])
    for comparison in payload.get("comparisons", []):
        if not isinstance(comparison, dict):
            continue
        lines.append(
            "| "
            + " | ".join(
                [
                    f"`{comparison.get('left_label', '')}`",
                    f"`{comparison.get('right_label', '')}`",
                    f"`{comparison.get('relation', '')}`",
                    f"`{comparison.get('common_commit_count', 0)}`",
                ]
            )
            + " |"
        )
    lines.extend(
        [
            "",
            "> 注意：本檔只比較本機 Git 物件，未 fetch 遠端；common commit sample 只用 SHA，不含 commit message。",
            "",
        ]
    )
    path.write_text("\n".join(lines), encoding="utf-8")


def main() -> int:
    parser = argparse.ArgumentParser()
    parser.add_argument("--group-name", required=True)
    parser.add_argument("--repo", action="append", type=parse_repo_arg, required=True)
    parser.add_argument("--sample-limit", type=int, default=5000)
    parser.add_argument("--git-timeout", type=int, default=10)
    parser.add_argument("--output-json", required=True)
    parser.add_argument("--output-md", required=True)
    args = parser.parse_args()

    payload = build_payload(args.group_name, args.repo, args.sample_limit, args.git_timeout)
    Path(args.output_json).write_text(
        json.dumps(payload, ensure_ascii=False, indent=2) + "\n",
        encoding="utf-8",
    )
    write_markdown(payload, Path(args.output_md))
    return 0


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