"""Build tools_index.yaml — classifies all tools/ Python files as active, support, or archive.

Active: called directly from spec/41_release_dag.yaml
Support: imported by other active tools (not entry-point but needed)
Archive: no direct or transitive reference from DAG or active tools
"""
from __future__ import annotations

import argparse
import re
from pathlib import Path
from typing import Any

import yaml


ROOT = Path(__file__).resolve().parents[1]
TOOLS_DIR = ROOT / "tools"
DAG_SPEC = ROOT / "spec" / "41_release_dag.yaml"
DEFAULT_OUT = ROOT / "tools_index.yaml"


def _load_dag_active(dag_path: Path) -> set[str]:
    if not dag_path.exists():
        return set()
    try:
        data = yaml.safe_load(dag_path.read_text(encoding="utf-8")) or {}
    except Exception:
        return set()
    active: set[str] = set()
    for node in (data.get("dag") or {}).get("nodes", {}).values():
        cmd = node.get("command", "")
        if isinstance(cmd, list):
            cmd = " ".join(str(c) for c in cmd)
        for m in re.findall(r"tools/([\w]+\.py)", str(cmd)):
            active.add(m)
    return active


def _find_imports(py_file: Path) -> set[str]:
    text = py_file.read_text(encoding="utf-8", errors="ignore")
    imports: set[str] = set()
    for m in re.findall(r"from\s+([\w_]+)\s+import|import\s+([\w_]+)", text):
        name = (m[0] or m[1]).strip()
        if name:
            candidate = f"{name}.py"
            imports.add(candidate)
    return imports


def _compute_support(active: set[str], all_tools: set[str]) -> set[str]:
    support: set[str] = set()
    frontier = set(active)
    for _ in range(5):
        new_support: set[str] = set()
        for tool_name in frontier:
            py = TOOLS_DIR / tool_name
            if not py.exists():
                continue
            for imp in _find_imports(py):
                if imp in all_tools and imp not in active and imp not in support:
                    new_support.add(imp)
        if not new_support:
            break
        support.update(new_support)
        frontier = new_support
    return support


def main() -> int:
    ap = argparse.ArgumentParser()
    ap.add_argument("--root", default=str(ROOT))
    ap.add_argument("--out", default=str(DEFAULT_OUT))
    args = ap.parse_args()

    tools_dir = Path(args.root) / "tools"
    all_tool_names = {p.name for p in tools_dir.glob("*.py")}

    dag_active = _load_dag_active(DAG_SPEC)
    support = _compute_support(dag_active, all_tool_names)

    active_list = sorted(dag_active & all_tool_names)
    support_list = sorted(support - dag_active)
    archive_candidates = sorted(all_tool_names - dag_active - support)

    result: dict[str, Any] = {
        "schema_version": "tools_index.v1",
        "generated_at": "2026-06-10",
        "summary": {
            "total": len(all_tool_names),
            "active": len(active_list),
            "support": len(support_list),
            "archive_candidates": len(archive_candidates),
        },
        "active": active_list,
        "support": support_list,
        "archive_candidates": archive_candidates,
    }

    out_path = Path(args.out)
    out_path.write_text(yaml.dump(result, allow_unicode=True, default_flow_style=False), encoding="utf-8")

    print("TOOL_INVENTORY_V1_OK")
    print(f"  total: {len(all_tool_names)}")
    print(f"  active (in DAG): {len(active_list)}")
    print(f"  support (imported by active): {len(support_list)}")
    print(f"  archive_candidates: {len(archive_candidates)}")
    return 0


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