#!/usr/bin/env python3
"""Build a deterministic inventory for the GMOCCAPY source and case assets."""

from __future__ import annotations

import hashlib
import json
from pathlib import Path


ROOT = Path(__file__).resolve().parents[1]
OUT = Path(__file__).resolve().parent


def sha256(path: Path) -> str:
    return hashlib.sha256(path.read_bytes()).hexdigest()


def classify(rel: str) -> str:
    path = Path(rel)
    if "/configs/sim/gmoccapy/" in rel:
        if path.suffix == ".ini":
            return "simulation_ini"
        if path.suffix in {".hal", ".tbl", ".para", ".var", ".pref"}:
            return "simulation_runtime_data"
        if path.suffix in {".ngc", ".py", ".ui", ".glade", ".svg", ".txt", ".README"}:
            return "simulation_resource"
        return "simulation_asset"
    if "/src/emc/usr_intf/gmoccapy/" in rel or "/lib/python/gmoccapy/" in rel:
        return "implementation"
    if "/share/gmoccapy/" in rel:
        if "icons/" in rel:
            return "icon_asset"
        return "installed_ui_asset"
    if "/src/po/gmoccapy/" in rel:
        return "translation"
    if "/docs/src/gui/images/gmoccapy_" in rel or "/docs/src/user/images/gmoccapy_" in rel:
        return "documentation_asset"
    if "/docs/src/gui/gmoccapy" in rel or "/docs/man/man1/gmoccapy" in rel:
        return "documentation"
    return "cross_reference"


def collect(paths: list[Path]) -> list[dict[str, object]]:
    files: list[Path] = []
    for base in paths:
        if base.is_file():
            files.append(base)
        elif base.is_dir():
            files.extend(p for p in base.rglob("*") if p.is_file())
    rows = []
    for path in sorted(set(files)):
        rel = path.relative_to(ROOT).as_posix()
        rows.append(
            {
                "path": rel,
                "role": classify(rel),
                "bytes": path.stat().st_size,
                "sha256": sha256(path),
            }
        )
    return rows


def main() -> None:
    rows = collect(
        [
            ROOT / "linuxcnc/configs/sim/gmoccapy",
            ROOT / "linuxcnc/src/emc/usr_intf/gmoccapy",
            ROOT / "linuxcnc/lib/python/gmoccapy",
            ROOT / "linuxcnc/share/gmoccapy",
            ROOT / "linuxcnc/src/po/gmoccapy",
            ROOT / "linuxcnc/docs/src/gui/gmoccapy.adoc",
            ROOT / "linuxcnc/docs/man/man1/gmoccapy.1",
            *sorted((ROOT / "linuxcnc/docs/src/gui/images").glob("gmoccapy_*")),
            *sorted((ROOT / "linuxcnc/docs/src/user/images").glob("gmoccapy_*")),
        ]
    )
    summary: dict[str, int] = {}
    for row in rows:
        role = str(row["role"])
        summary[role] = summary.get(role, 0) + 1
    payload = {
        "schema": "gmoccapy-file-inventory.v1",
        "scope": [
            "linuxcnc/configs/sim/gmoccapy/**",
            "linuxcnc/src/emc/usr_intf/gmoccapy/**",
            "linuxcnc/lib/python/gmoccapy/**",
            "linuxcnc/share/gmoccapy/**",
            "linuxcnc/src/po/gmoccapy/**",
            "linuxcnc/docs/src/gui/gmoccapy.adoc",
            "linuxcnc/docs/man/man1/gmoccapy.1",
            "linuxcnc/docs/src/gui/images/gmoccapy_*",
            "linuxcnc/docs/src/user/images/gmoccapy_*",
        ],
        "file_count": len(rows),
        "role_counts": dict(sorted(summary.items())),
        "files": rows,
    }
    (OUT / "gmoccapy-file-inventory.json").write_text(
        json.dumps(payload, ensure_ascii=False, sort_keys=True, indent=2) + "\n",
        encoding="utf-8",
    )
    lines = [
        "# GMOCCAPY 文件清单",
        "",
        f"清单范围内共 `{len(rows)}` 个文件；每项包含相对路径、角色、字节数和 SHA-256。",
        "",
        "| 角色 | 数量 |",
        "|---|---:|",
    ]
    lines.extend(f"| `{role}` | {count} |" for role, count in sorted(summary.items()))
    lines.extend(["", "完整逐文件数据见 [`gmoccapy-file-inventory.json`](gmoccapy-file-inventory.json)。", ""])
    (OUT / "gmoccapy-file-inventory.md").write_text("\n".join(lines), encoding="utf-8")


if __name__ == "__main__":
    main()