#!/usr/bin/env python3
"""
vCard deduplicator — conservative, correct.

Rules (in order):
  1. Same exact email address → same person
  2. Same phone (appears ≤2 times total — personal phones only) → same person
  3. Same normalized full name (≥2 significant words) → same person

Post-merge guard:
  If a merged record would have >4 different email domains → bad chain merge.
  Revert: keep each original record separately.
"""

import glob, re
from collections import defaultdict

INPUT_DIR = "/home/johan/clawd/tmp/contacts"
OUTPUT    = "/home/johan/clawd/tmp/contacts/merged.vcf"

# ── parser ────────────────────────────────────────────────────────────────────

def parse_vcf(path):
    with open(path, encoding='utf-8', errors='replace') as f:
        raw = re.sub(r'\r?\n[ \t]', '', f.read())
    blocks = [b for b in re.split(r'(?=BEGIN:VCARD)', raw, flags=re.I)
              if b.strip().upper().startswith('BEGIN:VCARD')]
    cards = []
    for block in blocks:
        card = defaultdict(list)
        for line in block.splitlines():
            if ':' not in line: continue
            k, _, v = line.partition(':')
            k = k.strip().upper(); v = v.strip()
            if k in ('BEGIN', 'END', 'VERSION'): continue
            card[k].append(v)
        cards.append(dict(card))
    return cards

# ── field helpers ─────────────────────────────────────────────────────────────

def get_field(card, prefix):
    for k, vs in card.items():
        if re.match(r'(ITEM\d+\.)?' + prefix + r'($|[;:])', k, re.I):
            return vs[0] if vs else ''
    return ''

def get_emails(card):
    out = set()
    for k, vs in card.items():
        if re.match(r'(ITEM\d+\.)?EMAIL', k, re.I):
            for v in vs:
                if v and '@' in v:
                    out.add(v.strip().lower())
    return out

def get_phones(card):
    out = {}  # norm → original
    for k, vs in card.items():
        if re.match(r'(ITEM\d+\.)?TEL', k, re.I):
            for v in vs:
                if v:
                    digits = re.sub(r'\D', '', v)
                    norm = digits[-9:] if len(digits) >= 9 else digits
                    if norm:
                        out[norm] = v.strip()
    return out

def get_name(card):
    fn = get_field(card, 'FN').strip()
    if not fn:
        fn = get_field(card, 'ORG').strip()
    return fn

def normalize_name(name):
    """Lowercase + sort words (order-independent matching)."""
    words = re.sub(r'\s+', ' ', name.strip().lower()).split()
    return ' '.join(sorted(words))

def completeness(card):
    return sum(len([v for v in vs if v]) for vs in card.values())

def email_domains(emails):
    return set(e.split('@')[1] for e in emails if '@' in e)

# ── merge ─────────────────────────────────────────────────────────────────────

NOISE_CATS = re.compile(
    r'^(imported on .+|restored from google.*|mycontacts)$', re.I)

def merge_cards(cards):
    base = max(cards, key=completeness)

    all_emails = set()
    all_phones = {}
    all_cats   = []
    for c in cards:
        all_emails |= get_emails(c)
        all_phones.update(get_phones(c))
        for k, vs in c.items():
            if k.upper() == 'CATEGORIES':
                for v in vs:
                    all_cats.extend(x.strip() for x in v.split(','))
    all_cats = list(dict.fromkeys(
        c for c in all_cats if not NOISE_CATS.match(c)))

    out = {}
    skip = ('EMAIL', 'TEL', 'CATEGORIES')
    for k, vs in base.items():
        if any(re.match(r'(ITEM\d+\.)?' + p, k, re.I) for p in skip):
            continue
        out[k] = list(vs)

    for i, email in enumerate(sorted(all_emails)):
        out[f'ITEM{i+1}.EMAIL;TYPE=INTERNET'] = [email]

    for i, (norm, orig) in enumerate(all_phones.items()):
        out[f'TEL_PHONE_{i}'] = [orig]

    if all_cats:
        out['CATEGORIES'] = [','.join(all_cats)]

    return out

def serialize(card):
    lines = ['BEGIN:VCARD', 'VERSION:3.0']
    priority = ['FN','N','ORG','TITLE','EMAIL','TEL','ADR','URL',
                'NOTE','BDAY','PHOTO','CATEGORIES','X-']
    def key_order(k):
        ku = k.upper()
        for i, p in enumerate(priority):
            if ku.startswith(p): return i
        return 99
    for k in sorted(card.keys(), key=key_order):
        display = re.sub(r'^TEL_PHONE_\d+$', 'TEL;TYPE=CELL', k)
        for v in card[k]:
            if v:
                lines.append(f'{display}:{v}')
    lines.append('END:VCARD')
    return '\n'.join(lines)

# ── union-find ────────────────────────────────────────────────────────────────

def make_uf(n):
    g = list(range(n))
    def find(x):
        while g[x] != x:
            g[x] = g[g[x]]; x = g[x]
        return x
    def union(a, b):
        ra, rb = find(a), find(b)
        if ra != rb: g[rb] = ra
    return find, union

# ── main ──────────────────────────────────────────────────────────────────────

def main():
    files = sorted(glob.glob(f'{INPUT_DIR}/*.vcf'))
    files = [f for f in files
             if not any(x in f for x in ('merged','final','dedup','postprocess'))]

    all_cards = []
    for f in files:
        cards = parse_vcf(f)
        print(f"  {f.split('/')[-1]}: {len(cards)}")
        all_cards.extend(cards)
    n = len(all_cards)
    print(f"Total: {n}")

    find, union = make_uf(n)

    # ── Rule 1: same email ────────────────────────────────────────────────────
    email_map = {}
    for i, c in enumerate(all_cards):
        for e in get_emails(c):
            if e in email_map:
                union(i, email_map[e])
            else:
                email_map[e] = i

    # ── Rule 2: same phone (personal only — skip phones in 3+ contacts) ───────
    phone_freq = defaultdict(int)
    for c in all_cards:
        for norm in get_phones(c):
            phone_freq[norm] += 1

    phone_map = {}
    for i, c in enumerate(all_cards):
        for norm, orig in get_phones(c).items():
            if phone_freq[norm] >= 3:
                continue  # shared/switchboard — skip
            if norm in phone_map:
                union(i, phone_map[norm])
            else:
                phone_map[norm] = i

    # ── Rule 3: exact full name (≥2 significant words) ────────────────────────
    name_map = {}
    for i, c in enumerate(all_cards):
        name = get_name(c)
        sig = [w for w in name.split() if len(w) > 2]
        if len(sig) < 2:
            continue  # single word / too short — skip
        key = normalize_name(name)
        if key in name_map:
            union(i, name_map[key])
        else:
            name_map[key] = i

    # ── Group ─────────────────────────────────────────────────────────────────
    groups = defaultdict(list)
    for i, c in enumerate(all_cards):
        groups[find(i)].append(c)

    # ── Merge + post-merge guard ──────────────────────────────────────────────
    MAX_DOMAINS = 4  # >4 email domains = chain merge gone wrong → revert

    result      = []
    reverted    = 0
    dup_removed = 0

    for root, grp in groups.items():
        if len(grp) == 1:
            result.append(grp[0])
            continue

        merged = merge_cards(grp)
        domains = email_domains(get_emails(merged))

        if len(domains) > MAX_DOMAINS:
            # Bad chain merge — keep each record individually
            reverted += len(grp) - 1
            for c in grp:
                result.append(c)
        else:
            dup_removed += len(grp) - 1
            result.append(merged)

    print(f"Duplicates removed: {dup_removed}")
    print(f"Reverted (chain merges): {reverted} groups → individual records")
    print(f"Output: {len(result)}")

    with open(OUTPUT, 'w', encoding='utf-8') as f:
        f.write('\n\n'.join(serialize(c) for c in result) + '\n')
    print(f"Written: {OUTPUT}")

if __name__ == '__main__':
    main()