#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
cleanup_cross_branch_duplicates.py

Limpieza generalizada de contactos en Marca que existen fisicamente en mas de
una sucursal (duplicados cruzados). Es la version generica del script puntual
cleanup_puebla_qro_duplicates.py.

Reglas de decision para elegir el lado a CONSERVAR:

  1. VALOR monetario: el lado con mayor suma de monetary_value en sus opps gana.
     - Si solo un lado tiene valor > 0, ese se conserva.
  2. TIENDA en Marca: si hay empate de valor (incluido el caso "ambos en $0"),
     se conserva el lado cuya sucursal coincide con el campo TIENDA del
     contacto Marca (resuelto via el verificador de sucursales).
  3. AMBIGUO: si ninguna regla resuelve, el caso se marca needs_review y NO
     se procesa automaticamente. Hay que correrlo con --include-ambiguous y
     explicitar --keeper-loc o --only-contact para forzar la decision manual.

Para cada par decidido:
  - En la(s) sucursal(es) PERDEDORA(S): elimina la opp residual y luego el
    contacto residual (orden importante para que el audit log mantenga el
    contact_id de la opp aunque GHL cascadee).
  - En MARCA: si la opp Marca tiene monetary_value o status distintos al
    lado ganador, hace PUT para sincronizar (la opp Marca pudo haberse
    quedado pegada a los datos del lado perdedor en la ultima sync).

Reglas de seguridad:
  - Dry-run por default. --apply para escribir.
  - Recalcula el plan desde DB en cada corrida (nunca consume JSON externo).
  - audit con run_id; los DELETE registran snapshot completo en old_value_json.
  - --exclude-contact ID  para excluir casos especificos (p.ej. miguel angel).
  - --only-contact ID     para procesar un solo caso.
  - --skip-pair "A,B"     para excluir todos los casos cuyo par sea exactamente
                          esas dos sucursales (case-insensitive, por nombre corto).

Uso:
    python scripts/cleanup_cross_branch_duplicates.py                          # dry-run de todo
    python scripts/cleanup_cross_branch_duplicates.py --apply --yes            # apply real
    python scripts/cleanup_cross_branch_duplicates.py --exclude-contact <id>   # excluir uno
    python scripts/cleanup_cross_branch_duplicates.py --only-contact <id>      # uno solo
    python scripts/cleanup_cross_branch_duplicates.py --skip-pair "Eugenia,Temixco"
"""

import argparse
import json
import os
import sqlite3
import sys
import uuid
from collections import defaultdict
from datetime import datetime

ROOT_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
if ROOT_DIR not in sys.path:
    sys.path.insert(0, ROOT_DIR)

from scripts.audit_brand_vs_branches_totals import (  # noqa: E402
    load_accounts_filtered, load_contacts, load_opps,
    build_contact_index, find_match,
    extract_tienda_from_custom_fields, resolve_tienda_field_id,
    load_verifier, normalize_tienda,
    BRAND_LOCATION_ID, DB_PATH,
)
import sync_engine  # noqa: E402
import script_audit  # noqa: E402

SCRIPT_NAME = "cleanup_cross_branch_duplicates.py"


def safe_print(*args, **kwargs):
    sep = kwargs.get("sep", " ")
    end = kwargs.get("end", "\n")
    text = sep.join(str(a) for a in args)
    encoding = sys.stdout.encoding or "utf-8"
    try:
        sys.stdout.write(text + end)
        sys.stdout.flush()
    except UnicodeEncodeError:
        sys.stdout.write(text.encode(encoding, errors="replace").decode(encoding) + end)
        sys.stdout.flush()


def _short_branch_name(full):
    """Devuelve la parte despues del ultimo '-' del nombre de la sucursal."""
    return (full or "").split("-")[-1].strip()


def _norm_branch(name):
    """Normalizacion para comparar pares ignorando acentos/case."""
    s = (name or "").lower()
    return (s.replace("é", "e").replace("á", "a").replace("í", "i")
             .replace("ó", "o").replace("ú", "u").strip())


def build_plan(log=safe_print, skip_pairs=None, exclude_contacts=None, only_contacts=None,
               include_ambiguous=False):
    """Recalcula el plan completo. Devuelve dict:
      {
        decisions: [...],            # cada item con keeper + losers + criterio
        ambiguous: [...],            # casos sin regla clara (excluidos del apply)
        excluded_by_filter: [...],   # casos saltados por skip_pairs/exclude_contacts/only_contacts
      }
    """
    if not os.path.exists(DB_PATH):
        raise FileNotFoundError(f"No existe {DB_PATH}. Sincroniza primero.")

    skip_pairs_norm = set()
    for sp in skip_pairs or []:
        parts = [_norm_branch(p) for p in sp.split(",")]
        if len(parts) == 2:
            skip_pairs_norm.add(tuple(sorted(parts)))
    exclude_set = set(exclude_contacts or [])
    only_set = set(only_contacts or [])

    conn = sqlite3.connect(DB_PATH)
    conn.row_factory = sqlite3.Row
    try:
        _brand, branches, _ = load_accounts_filtered(conn)
        brand_contacts = load_contacts(conn, BRAND_LOCATION_ID)
        brand_opps = load_opps(conn, BRAND_LOCATION_ID)
        brand_tienda_field_id = resolve_tienda_field_id(conn, BRAND_LOCATION_ID)
        verifier_by_loc, verifier_by_tienda = load_verifier()

        branch_idx = {}
        branch_name_by_loc = {}
        branch_opps_by_cid = {}
        branch_contacts_by_loc = {}
        for b in branches:
            loc = b["location_id"]
            branch_name_by_loc[loc] = b["nombre"]
            bc = load_contacts(conn, loc)
            bo = load_opps(conn, loc)
            branch_idx[loc] = build_contact_index(bc)
            branch_contacts_by_loc[loc] = {c["id"]: c for c in bc}
            grp = defaultdict(list)
            for o in bo:
                grp[o["contact_id"]].append(o)
            branch_opps_by_cid[loc] = grp

        brand_by_id = {c["id"]: c for c in brand_contacts}
        brand_opps_by_cid_marca = defaultdict(list)
        for o in brand_opps:
            brand_opps_by_cid_marca[o["contact_id"]].append(o)

        contact_branches = defaultdict(list)
        phone_collisions_per_brand = defaultdict(list)
        for c in brand_contacts:
            for loc in branch_idx:
                idxp, idxe, idxn = branch_idx[loc]
                matches, collisions = find_match(
                    c, idxp, idxe, idxn, return_collisions=True
                )
                for m in matches:
                    contact_branches[c["id"]].append((loc, m["id"]))
                for m in collisions:
                    phone_collisions_per_brand[c["id"]].append((loc, m["id"]))

        decisions = []
        ambiguous = []
        excluded_by_filter = []
        # Colisiones de teléfono: contactos Marca con phone idéntico al de
        # algún contacto en sucursal pero con nombre divergente. No los
        # tratamos como duplicados (probablemente son personas distintas
        # compartiendo número, ej. una pareja). Se reportan aparte.
        phone_collisions = []
        for bid, hits in phone_collisions_per_brand.items():
            if not hits:
                continue
            c = brand_by_id.get(bid)
            if not c:
                continue
            phone_collisions.append({
                "brand_contact_id": bid,
                "name": (f"{c.get('first_name') or ''} {c.get('last_name') or ''}".strip() or "(sin nombre)"),
                "phone": c.get("phone") or "",
                "colliding_branches": [
                    {
                        "location_id": loc,
                        "location_name": branch_name_by_loc.get(loc, loc),
                        "branch_contact_id": bcid,
                    }
                    for loc, bcid in hits
                ],
            })

        for bid, locs in contact_branches.items():
            distinct_locs = {l[0] for l in locs}
            if len(distinct_locs) < 2:
                continue  # intra-sucursal no aplica aqui (otro script)

            c = brand_by_id[bid]
            name = (f"{c.get('first_name') or ''} {c.get('last_name') or ''}".strip() or "(sin nombre)")
            tienda_raw = extract_tienda_from_custom_fields(c.get("custom_fields_json"), brand_tienda_field_id) or ""
            tienda_norm = normalize_tienda(tienda_raw) if tienda_raw else ""
            tienda_loc = verifier_by_tienda.get(tienda_norm)

            # Filtros
            if only_set and bid not in only_set:
                continue
            if bid in exclude_set:
                excluded_by_filter.append({"brand_contact_id": bid, "name": name, "reason": "exclude_contact"})
                continue

            # Construir candidatos por sucursal
            per = []
            for loc, bcid in locs:
                opps = branch_opps_by_cid[loc].get(bcid, [])
                per.append({
                    "loc": loc,
                    "bcid": bcid,
                    "short_name": _short_branch_name(branch_name_by_loc[loc]),
                    "full_name": branch_name_by_loc[loc],
                    "n_opps": len(opps),
                    "total_val": sum(float(o.get("monetary_value") or 0) for o in opps),
                    "opps": opps,
                })

            # Skip-pair filter (por nombre normalizado de las sucursales del par)
            pair_norm = tuple(sorted({_norm_branch(p["short_name"]) for p in per}))
            if len(pair_norm) == 2 and pair_norm in skip_pairs_norm:
                excluded_by_filter.append({
                    "brand_contact_id": bid, "name": name,
                    "reason": f"skip_pair {pair_norm[0]}<->{pair_norm[1]}",
                })
                continue

            # Regla 1: VALOR
            by_val = sorted(per, key=lambda x: -x["total_val"])
            max_val = by_val[0]["total_val"]
            candidates_max = [p for p in per if p["total_val"] == max_val]
            keeper = None
            criterio = None
            if max_val > 0 and len(candidates_max) == 1:
                keeper = candidates_max[0]
                criterio = "VALOR"
            elif max_val > 0 and len(candidates_max) > 1:
                # Empate de valor -> TIENDA decide
                match = [p for p in candidates_max if p["loc"] == tienda_loc]
                if len(match) == 1:
                    keeper = match[0]
                    criterio = "VAL+TIENDA"
            else:
                # Todos en $0 -> TIENDA decide
                match = [p for p in per if p["loc"] == tienda_loc]
                if len(match) == 1:
                    keeper = match[0]
                    criterio = "TIENDA"

            if not keeper:
                amb = {
                    "brand_contact_id": bid, "name": name, "tienda": tienda_raw,
                    "candidates": [{"loc": p["loc"], "name": p["short_name"],
                                    "n_opps": p["n_opps"], "total_val": p["total_val"]} for p in per],
                }
                ambiguous.append(amb)
                continue

            losers = [p for p in per if p["loc"] != keeper["loc"]]

            # Plan de delete en sucursales perdedoras
            loser_deletions = []
            for l in losers:
                loser_deletions.append({
                    "location_id": l["loc"],
                    "location_name": l["full_name"],
                    "contact_id": l["bcid"],
                    "contact_snapshot": {
                        "id": l["bcid"],
                        "first_name": branch_contacts_by_loc[l["loc"]][l["bcid"]].get("first_name"),
                        "last_name": branch_contacts_by_loc[l["loc"]][l["bcid"]].get("last_name"),
                        "phone": branch_contacts_by_loc[l["loc"]][l["bcid"]].get("phone"),
                        "email": branch_contacts_by_loc[l["loc"]][l["bcid"]].get("email"),
                    },
                    "opps_snapshots": [
                        {"id": o["id"], "name": o.get("name"), "status": o.get("status"),
                         "monetary_value": o.get("monetary_value"),
                         "pipeline_id": o.get("pipeline_id"),
                         "pipeline_stage_id": o.get("pipeline_stage_id")}
                        for o in l["opps"]
                    ],
                })

            # Plan de update en Marca: si la opp Marca difiere del keeper en valor/status
            brand_opp_updates = []
            m_opps = brand_opps_by_cid_marca.get(bid, [])
            k_opps = keeper["opps"]
            if m_opps and k_opps:
                m_o = m_opps[0]
                k_o = k_opps[0]
                m_val = float(m_o.get("monetary_value") or 0)
                k_val = float(k_o.get("monetary_value") or 0)
                m_st = (m_o.get("status") or "").lower()
                k_st = (k_o.get("status") or "").lower()
                if m_val != k_val or m_st != k_st:
                    brand_opp_updates.append({
                        "brand_opp_id": m_o["id"],
                        "old": {"monetary_value": m_val, "status": m_st},
                        "new": {"monetary_value": k_val, "status": k_st},
                    })

            decisions.append({
                "brand_contact_id": bid,
                "name": name,
                "tienda": tienda_raw,
                "criterio": criterio,
                "keeper": {
                    "location_id": keeper["loc"],
                    "location_name": keeper["full_name"],
                    "short_name": keeper["short_name"],
                    "total_val": keeper["total_val"],
                },
                "losers": loser_deletions,
                "brand_opp_updates": brand_opp_updates,
            })

        return {
            "decisions": decisions,
            "ambiguous": ambiguous,
            "excluded_by_filter": excluded_by_filter,
            "phone_collisions": phone_collisions,
        }
    finally:
        conn.close()


def _empty_summary():
    return {
        "decisions": 0,
        "ambiguous_skipped": 0,
        "filter_excluded": 0,
        "phone_collisions_unresolved": 0,
        "loser_contacts_to_delete": 0,
        "loser_opps_to_delete": 0,
        "brand_opps_to_update": 0,
        "loser_contacts_deleted": 0,
        "loser_opps_deleted": 0,
        "brand_opps_updated": 0,
        "errors": 0,
    }


def run_cleanup(dry_run=True, log=None, run_id=None,
                skip_pairs=None, exclude_contacts=None, only_contacts=None,
                include_ambiguous=False):
    if log is None:
        log = safe_print

    plan = build_plan(
        log=log, skip_pairs=skip_pairs,
        exclude_contacts=exclude_contacts, only_contacts=only_contacts,
        include_ambiguous=include_ambiguous,
    )

    summary = _empty_summary()
    summary["decisions"] = len(plan["decisions"])
    summary["ambiguous_skipped"] = len(plan["ambiguous"])
    summary["filter_excluded"] = len(plan["excluded_by_filter"])
    summary["phone_collisions_unresolved"] = len(plan.get("phone_collisions") or [])
    summary["loser_contacts_to_delete"] = sum(len(d["losers"]) for d in plan["decisions"])
    summary["loser_opps_to_delete"] = sum(len(l["opps_snapshots"]) for d in plan["decisions"] for l in d["losers"])
    summary["brand_opps_to_update"] = sum(len(d["brand_opp_updates"]) for d in plan["decisions"])

    log(f"[{datetime.now().strftime('%H:%M:%S')}] === cleanup_cross_branch_duplicates ===")
    log(f"Modo: {'DRY-RUN (no escribe)' if dry_run else 'APPLY (escribe en GHL)'}")
    log(f"Decisiones automaticas: {summary['decisions']} | ambiguos saltados: {summary['ambiguous_skipped']} | filtrados: {summary['filter_excluded']}")
    log(f"A eliminar en sucursales perdedoras: {summary['loser_contacts_to_delete']} contactos, {summary['loser_opps_to_delete']} opps")
    log(f"A actualizar en Marca: {summary['brand_opps_to_update']} opps")
    if summary["phone_collisions_unresolved"]:
        log(f"Colisiones telefono sin match por nombre (NO se procesan): {summary['phone_collisions_unresolved']}")

    if plan.get("phone_collisions"):
        log("\nColisiones de telefono (mismo numero, distinto nombre - revision manual):")
        for col in plan["phone_collisions"]:
            branches_str = ", ".join(
                f"{cb['location_name']}:{cb['branch_contact_id']}" for cb in col["colliding_branches"]
            )
            log(f"  - Marca {col['name']:<30} tel={col['phone']:<14} contra: {branches_str}")

    if plan["ambiguous"]:
        log("\nAmbiguos (NO se procesan):")
        for a in plan["ambiguous"]:
            cands_str = " vs ".join(f"{c['name']}({c['n_opps']}/${c['total_val']:,.0f})" for c in a["candidates"])
            log(f"  - {a['name']:<30}  TIENDA={a['tienda']}  candidatos: {cands_str}")

    if plan["excluded_by_filter"]:
        log("\nExcluidos por filtro:")
        for e in plan["excluded_by_filter"]:
            log(f"  - {e['name']}  ({e['reason']})")

    if not dry_run:
        tokens_map = sync_engine.get_tokens_map()
        brand_token = tokens_map.get(BRAND_LOCATION_ID)
        if not brand_token:
            raise RuntimeError(f"No hay token para Marca {BRAND_LOCATION_ID}")
        client = sync_engine.ghl_client

    items = []
    for d in plan["decisions"]:
        item = {
            "brand_contact_id": d["brand_contact_id"],
            "name": d["name"],
            "criterio": d["criterio"],
            "keeper": d["keeper"]["short_name"],
            "deletions": [],
            "brand_updates": [],
            "status": "pending",
            "error": None,
        }
        try:
            # ---- Borrar en sucursales perdedoras ----
            for loser in d["losers"]:
                loc = loser["location_id"]
                loc_name = loser["location_name"]
                short = _short_branch_name(loc_name)
                if not dry_run:
                    loser_token = tokens_map.get(loc)
                    if not loser_token:
                        raise RuntimeError(f"No hay token para {loc_name} ({loc})")

                # Opps de la perdedora
                for opp_snap in loser["opps_snapshots"]:
                    opp_id = opp_snap["id"]
                    if dry_run:
                        item["deletions"].append({"loc": short, "opp_id": opp_id, "status": "would_delete"})
                        continue
                    cid = script_audit.record_change(
                        run_id, loc, "opportunity", opp_id,
                        "", "deleted_residual_dup", opp_snap, None,
                    ) if run_id else None
                    try:
                        client.delete_opportunity(loser_token, opp_id, loc)
                        summary["loser_opps_deleted"] += 1
                        item["deletions"].append({"loc": short, "opp_id": opp_id, "status": "deleted"})
                        if cid: script_audit.mark_change(cid, "applied")
                    except Exception as e:
                        summary["errors"] += 1
                        item["deletions"].append({"loc": short, "opp_id": opp_id, "status": "error", "error": str(e)})
                        if cid: script_audit.mark_change(cid, "failed", error_message=str(e))
                        raise

                # Contacto perdedor
                if dry_run:
                    item["deletions"].append({"loc": short, "contact_id": loser["contact_id"], "status": "would_delete"})
                else:
                    cid = script_audit.record_change(
                        run_id, loc, "contact", loser["contact_id"],
                        "", "deleted_residual_dup", loser["contact_snapshot"], None,
                    ) if run_id else None
                    try:
                        client.delete_contact(loser_token, loser["contact_id"], loc)
                        summary["loser_contacts_deleted"] += 1
                        item["deletions"].append({"loc": short, "contact_id": loser["contact_id"], "status": "deleted"})
                        if cid: script_audit.mark_change(cid, "applied")
                    except Exception as e:
                        summary["errors"] += 1
                        item["deletions"].append({"loc": short, "contact_id": loser["contact_id"], "status": "error", "error": str(e)})
                        if cid: script_audit.mark_change(cid, "failed", error_message=str(e))
                        raise

            # ---- Actualizar opps Marca con datos del keeper ----
            for u in d["brand_opp_updates"]:
                if dry_run:
                    item["brand_updates"].append({"opp_id": u["brand_opp_id"], "status": "would_update", **u})
                else:
                    cid = script_audit.record_change(
                        run_id, BRAND_LOCATION_ID, "opportunity", u["brand_opp_id"],
                        "monetary_value+status", "updated_from_keeper",
                        u["old"], u["new"],
                    ) if run_id else None
                    try:
                        payload = {"monetaryValue": u["new"]["monetary_value"]}
                        client.update_opportunity(brand_token, u["brand_opp_id"], payload)
                        if u["old"]["status"] != u["new"]["status"]:
                            client.update_opportunity_status(brand_token, u["brand_opp_id"], u["new"]["status"])
                        summary["brand_opps_updated"] += 1
                        item["brand_updates"].append({"opp_id": u["brand_opp_id"], "status": "updated", **u})
                        if cid: script_audit.mark_change(cid, "applied")
                    except Exception as e:
                        summary["errors"] += 1
                        item["brand_updates"].append({"opp_id": u["brand_opp_id"], "status": "error", "error": str(e), **u})
                        if cid: script_audit.mark_change(cid, "failed", error_message=str(e))
                        raise

            item["status"] = "ok"
            log(f"  [{'DRY' if dry_run else 'OK'}] {d['name']:<30} keeper={d['keeper']['short_name']:<12} "
                f"criterio={d['criterio']:<11} losers={','.join(_short_branch_name(l['location_name']) for l in d['losers'])}")
        except Exception as e:
            item["status"] = "error"
            item["error"] = str(e)
            log(f"  [ERROR] {d['name']}: {e}")

        items.append(item)

    log("\n=== RESUMEN ===")
    for k, v in summary.items():
        log(f"  {k:<32}: {v}")

    return {
        "dry_run": dry_run,
        "summary": summary,
        "items": items,
        "ambiguous": plan["ambiguous"],
        "excluded_by_filter": plan["excluded_by_filter"],
        "phone_collisions": plan.get("phone_collisions") or [],
    }


def main():
    parser = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
    parser.add_argument("--apply", action="store_true", help="Ejecuta escrituras en GHL. Default dry-run.")
    parser.add_argument("--yes", action="store_true", help="Skip confirmacion interactiva.")
    parser.add_argument("--only-contact", action="append", default=[], help="Procesa solo el brand_contact_id dado.")
    parser.add_argument("--exclude-contact", action="append", default=[], help="Excluye brand_contact_id (repetible).")
    parser.add_argument("--skip-pair", action="append", default=[], help='Excluye casos del par "SucA,SucB" (nombre corto).')
    parser.add_argument("--include-ambiguous", action="store_true", help="Procesa tambien los ambiguos (NO recomendado sin --only-contact).")
    parser.add_argument("--json", action="store_true", help="Imprime resultado como JSON.")
    parser.add_argument("--run-id", type=str, default=None, help="Id de script_audit existente.")
    args = parser.parse_args()

    dry_run = not args.apply
    if not dry_run and not args.yes:
        safe_print("\nEsto eliminara contactos+opps en sucursales perdedoras y actualizara opps en Marca.")
        safe_print("Los DELETE son destructivos (audit log guarda snapshots para reconstruccion manual).")
        confirm = input("Continuar? (y/N): ").strip().lower()
        if confirm not in ("y", "yes", "s", "si", "sí"):
            safe_print("Cancelado.")
            return 1

    run_id = args.run_id
    if not dry_run:
        if not run_id:
            run_id = f"ccbd-{uuid.uuid4().hex[:12]}"
        try:
            script_audit.init_audit_db()
            script_audit.create_run(
                run_id, SCRIPT_NAME,
                arguments=f"--apply filters: only={args.only_contact or 'all'} exclude={args.exclude_contact or 'none'} skip_pairs={args.skip_pair or 'none'}",
                locations=[BRAND_LOCATION_ID],
                execution_mode="sequential",
            )
        except Exception as e:
            safe_print(f"[warn] no se pudo iniciar audit run: {e}")
            run_id = None

    try:
        result = run_cleanup(
            dry_run=dry_run, log=safe_print, run_id=run_id,
            skip_pairs=args.skip_pair or None,
            exclude_contacts=args.exclude_contact or None,
            only_contacts=args.only_contact or None,
            include_ambiguous=args.include_ambiguous,
        )
    except Exception as e:
        if run_id:
            try: script_audit.update_run_status(run_id, "failed", str(e))
            except Exception: pass
        safe_print(f"[FATAL] {e}")
        return 2

    if run_id:
        try:
            errors = result["summary"]["errors"]
            success_any = (result["summary"]["loser_contacts_deleted"]
                           + result["summary"]["loser_opps_deleted"]
                           + result["summary"]["brand_opps_updated"]) > 0
            status = "failed" if errors and not success_any else "success"
            script_audit.update_run_status(run_id, status)
        except Exception:
            pass
        result["run_id"] = run_id

    if args.json:
        safe_print(json.dumps(result, default=str, ensure_ascii=False, indent=2))
    return 0 if result["summary"]["errors"] == 0 else 1


if __name__ == "__main__":
    sys.exit(main() or 0)