#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Script: tests/compare_vaults.py
Version: 1.0.0 
Datum: 2025-09-10


Funktion
--------
Vergleicht zwei Ordner mit Markdown-Dateien (Vault vs. Export). Fokus:
  - body: reiner Body-Text (Whitespace tolerant)
  - frontmatter: YAML-Felder selektiv (id, title, type, tags, status, aliases)
  - both: erst FM, dann Body

Aufrufe
-------
    python3 tests/compare_vaults.py --src ./test_vault --dst ./_exportVault --focus body
    python3 tests/compare_vaults.py --src ./test_vault --dst ./_exportVault --focus both
"""
from __future__ import annotations

import argparse, os, sys, glob
from typing import Tuple, Dict, Any
import yaml

def split_md(p: str) -> Tuple[Dict[str, Any], str]:
    with open(p, "r", encoding="utf-8") as f:
        s = f.read()
    if s.startswith("---"):
        try:
            fm_txt, body = s.split("\n---\n", 1)
            fm = yaml.safe_load(fm_txt.strip("- \n")) or {}
        except Exception:
            fm, body = {}, s
    else:
        fm, body = {}, s
    return fm, body.strip()

def norm_body(s: str) -> str:
    return "\n".join([ln.rstrip() for ln in s.strip().splitlines()]).strip()

def main():
    ap = argparse.ArgumentParser()
    ap.add_argument("--src", required=True, help="Original-Vault")
    ap.add_argument("--dst", required=True, help="Export-Ordner")
    ap.add_argument("--focus", choices=["body","frontmatter","both"], default="body")
    args = ap.parse_args()

    src = os.path.abspath(args.src)
    dst = os.path.abspath(args.dst)

    # Map per Note-ID
    by_id = {}
    for p in glob.glob(os.path.join(src, "**", "*.md"), recursive=True):
        fm, body = split_md(p)
        nid = fm.get("id") or os.path.splitext(os.path.basename(p))[0]
        by_id.setdefault(nid, {})["src"] = (p, fm, body)
    for p in glob.glob(os.path.join(dst, "**", "*.md"), recursive=True):
        fm, body = split_md(p)
        nid = fm.get("id") or os.path.splitext(os.path.basename(p))[0]
        by_id.setdefault(nid, {})["dst"] = (p, fm, body)

    mismatches = 0
    for nid, d in sorted(by_id.items()):
        src_t = d.get("src")
        dst_t = d.get("dst")
        if not src_t or not dst_t:
            print({"note_id": nid, "status": "missing", "src": bool(src_t), "dst": bool(dst_t)})
            mismatches += 1
            continue
        sp, sfm, sbody = src_t
        dp, dfm, dbody = dst_t

        # frontmatter compare (subset)
        fm_ok = True
        fm_keys = ["id","title","type","status","tags","aliases"]
        if args.focus in ("frontmatter","both"):
            for k in fm_keys:
                if (sfm.get(k) or None) != (dfm.get(k) or None):
                    fm_ok = False
                    break
        # body compare
        body_ok = True
        if args.focus in ("body","both"):
            if norm_body(sbody) != norm_body(dbody):
                body_ok = False

        if not (fm_ok and body_ok):
            mismatches += 1
            print({
                "note_id": nid,
                "frontmatter_equal": fm_ok,
                "body_equal": body_ok,
                "src_path": sp,
                "dst_path": dp
            })

    if mismatches:
        print({"summary": "DIFFS", "count": mismatches})
        sys.exit(1)
    else:
        print({"summary": "OK", "count": 0})

if __name__ == "__main__":
    main()