"""
===============================================================================
FILE: services/scud_etl/merger.py
ROLE: Агрегация реестров, выбор совместителей 1С по отделу СКУД и авто-связки.
      Распределение исключений в общий рабочий пул при отсутствии справок.
===============================================================================
"""

import logging
from typing import Dict, Any, List
import pandas as pd

from services.knowledge.service import get_department_synonyms_dict
from config import normalize_fio, load_exceptions
from core.connection import get_connection

logger = logging.getLogger("SCUD_MERGER")


def load_identity_mappings() -> Dict[str, str]:
    with get_connection() as conn:
        cursor = conn.cursor()
        cursor.execute("""
            CREATE TABLE IF NOT EXISTS person_identity_mapping (
                id INTEGER PRIMARY KEY AUTOINCREMENT,
                scud_fio TEXT NOT NULL,
                zup_fio TEXT NOT NULL,
                scud_dept TEXT,
                zup_dept TEXT,
                match_source TEXT DEFAULT 'AI',
                status TEXT DEFAULT 'ACTIVE',
                confidence REAL DEFAULT 1.0,
                created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
                UNIQUE(scud_fio, zup_fio)
            );
        """)
        cursor.execute("""
            SELECT scud_fio, zup_fio 
            FROM person_identity_mapping 
            WHERE status = 'ACTIVE'
        """)
        return {r[0]: r[1] for r in cursor.fetchall()}


def aggregate_scud_by_person(df_scud: pd.DataFrame) -> pd.DataFrame:
    if df_scud is None or df_scud.empty:
        return df_scud

    df = df_scud.copy()
    if 'fio_clean' not in df.columns:
        fio_col = 'Сотрудник' if 'Сотрудник' in df.columns else 'ФИО'
        df['fio_clean'] = df[fio_col].apply(normalize_fio)

    mapping_dict = load_identity_mappings()
    if mapping_dict:
        df['fio_clean'] = df['fio_clean'].apply(lambda f: mapping_dict.get(f, f))

    aggregated_rows = []
    for fio_clean, group in df.groupby('fio_clean'):
        if len(group) == 1:
            aggregated_rows.append(group.iloc[0].to_dict())
            continue

        base_row = group.sort_values(by='Пришел', ascending=False).iloc[0].to_dict()

        valid_ins = [
            str(t).strip() for t in group['Начало_дня'] 
            if str(t).strip() not in ['Нет входа', '—', '', 'nan', 'None', '00:00:00', '00:00']
        ]
        base_row['Начало_дня'] = min(valid_ins) if valid_ins else 'Нет входа'

        valid_outs = [
            str(t).strip() for t in group['Конец_дня'] 
            if str(t).strip() not in ['Нет выхода', '—', '', 'nan', 'None', '00:00:00', '00:00']
        ]
        base_row['Конец_дня'] = max(valid_outs) if valid_outs else 'Нет выхода'

        valid_acts = [
            str(t).strip() for t in group['Первая_активность'] 
            if str(t).strip() not in ['—', '', 'nan', 'None', '00:00:00']
        ]
        base_row['Первая_активность'] = min(valid_acts) if valid_acts else '—'
        base_row['Пришел'] = any(group['Пришел'] == True) or (base_row['Начало_дня'] != 'Нет входа')

        durations = [str(d) for d in group['Находился_в_здании'] if str(d) not in ['00:00', '', 'nan']]
        if durations:
            base_row['Находился_в_здании'] = max(durations)

        aggregated_rows.append(base_row)

    return pd.DataFrame(aggregated_rows)


def select_best_zup_position(df_staff_1c: pd.DataFrame, df_scud_agg: pd.DataFrame) -> pd.DataFrame:
    if df_staff_1c is None or df_staff_1c.empty:
        return pd.DataFrame()

    df_staff = df_staff_1c.copy()
    if 'fio_clean' not in df_staff.columns:
        f_col = 'ФИО' if 'ФИО' in df_staff.columns else 'Сотрудник'
        df_staff['fio_clean'] = df_staff[f_col].apply(normalize_fio)

    scud_dept_map = {}
    if df_scud_agg is not None and not df_scud_agg.empty:
        for _, r in df_scud_agg.iterrows():
            scud_dept_map[r.get('fio_clean', '')] = str(r.get('Подразделение', '')).strip().upper()

    best_rows = []
    for fio, group in df_staff.groupby('fio_clean'):
        if len(group) == 1:
            best_rows.append(group.iloc[0].to_dict())
            continue

        target_scud_dept = scud_dept_map.get(fio, "")
        matched_row = None

        if target_scud_dept:
            for _, r in group.iterrows():
                dept_1c = str(r.get('Подразделение', '')).strip().upper()
                if dept_1c == target_scud_dept or target_scud_dept in dept_1c or dept_1c in target_scud_dept:
                    matched_row = r.to_dict()
                    break

        if not matched_row:
            matched_row = group.iloc[0].to_dict()

        best_rows.append(matched_row)

    return pd.DataFrame(best_rows)


def merge_scud_and_1c(
    df_scud: pd.DataFrame, 
    df_staff_1c: pd.DataFrame, 
    df_absences_1c: pd.DataFrame
) -> pd.DataFrame:
    if (df_scud is None or df_scud.empty) and (df_staff_1c is None or df_staff_1c.empty):
        return pd.DataFrame(columns=[
            'Сотрудник', 'fio_clean', 'Подразделение', 'Должность', 
            'Начало_дня', 'Первая_активность', 'Конец_дня', 'Находился_в_здании', 
            'Пришел', 'anomaly_flag', 'причина отсутствия', 'Вид_отсутствия', 'is_excluded'
        ])

    synonyms = get_department_synonyms_dict()
    exceptions_cfg = load_exceptions()

    df_scud_agg = aggregate_scud_by_person(df_scud)
    df_staff_agg = select_best_zup_position(df_staff_1c, df_scud_agg)

    df_res = df_scud_agg.copy() if df_scud_agg is not None and not df_scud_agg.empty else df_staff_agg.copy()

    if "Сотрудник" in df_res.columns:
        df_res["fio_clean"] = df_res["Сотрудник"].apply(normalize_fio)
    elif "ФИО" in df_res.columns:
        df_res["Сотрудник"] = df_res["ФИО"]
        df_res["fio_clean"] = df_res["ФИО"].apply(normalize_fio)
    elif "fio_clean" not in df_res.columns:
        df_res["fio_clean"] = ""

    for col, default_val in [
        ('Начало_дня', 'Нет входа'),
        ('Первая_активность', '—'),
        ('Конец_дня', 'Нет выхода'),
        ('Находился_в_здании', '00:00'),
        ('Пришел', False),
        ('anomaly_flag', 'NONE')
    ]:
        if col not in df_res.columns:
            df_res[col] = default_val

    reverse_synonyms = {v.lower(): k.upper() for k, v in synonyms.items()}
    direct_synonyms = {k.lower(): k.upper() for k in synonyms.keys()}
    all_dept_map = {**reverse_synonyms, **direct_synonyms, "отдел внутреннего контроля": "ОВК", "отдел вневедомственного контроля": "ОВК"}

    if "Подразделение" in df_res.columns:
        df_res["Подразделение"] = df_res["Подразделение"].apply(
            lambda d: all_dept_map.get(str(d).strip().lower(), str(d).strip())
        )

    absences_map = {}
    if df_absences_1c is not None and not df_absences_1c.empty:
        fio_col = next((c for c in ["fio_clean", "ФИО", "Сотрудник"] if c in df_absences_1c.columns), None)
        reason_col = next((c for c in ["Вид_отсутствия", "Причина", "причина отсутствия"] if c in df_absences_1c.columns), None)

        if fio_col and reason_col:
            for _, row in df_absences_1c.iterrows():
                fio = normalize_fio(str(row[fio_col]))
                reason = str(row[reason_col]).strip()
                if reason and reason.lower() != "nan":
                    absences_map[fio] = reason

    manual_reasons_map = {}
    try:
        from services.manual_absences_repo import get_active_manual_absences_for_date
        # date_clean берется из даты контекста либо из текущих суток
        target_date_val = df_res.get('Дата', pd.Series()).iloc[0] if 'Дата' in df_res.columns and not df_res.empty else None
        if target_date_val:
            m_records = get_active_manual_absences_for_date(str(target_date_val))
            for mr in m_records:
                manual_reasons_map[mr['fio_clean']] = mr['reason']
    except Exception:
        pass

    df_res["detailed_reason"] = df_res["fio_clean"].map(manual_reasons_map).fillna("")
    df_res["причина отсутствия"] = df_res["fio_clean"].map(absences_map)
    df_res["Вид_отсутствия"] = df_res["причина отсутствия"]

    exc_fios = [normalize_fio(f) for f in exceptions_cfg.get("fio", []) if f]
    exc_depts = [d.strip().upper() for d in exceptions_cfg.get("departments", []) if d]
    exc_pos = [p.strip().lower() for p in exceptions_cfg.get("positions", []) if p]
    pos_kw = [k.strip().lower() for k in exceptions_cfg.get("position_keywords", []) if k]
    whitelist_fios = [normalize_fio(f) for f in exceptions_cfg.get("include_fio", []) if f]

    df_res["is_excluded"] = False
    for idx, row in df_res.iterrows():
        fio = row.get("fio_clean", "")
        has_official_absence = pd.notna(row.get("Вид_отсутствия")) and str(row.get("Вид_отсутствия")).strip() not in ["", "nan", "None", "Исключение"]

        if fio in whitelist_fios:
            df_res.at[idx, "is_excluded"] = False
            continue

        dep = str(row.get("Подразделение", "")).upper()
        pos = str(row.get("Должность", "")).lower()

        is_match_exc = (fio in exc_fios or dep in exc_depts or any(d in dep for d in exc_depts) or pos in exc_pos or any(k in pos for k in pos_kw))

        if is_match_exc:
            if has_official_absence:
                df_res.at[idx, "is_excluded"] = False
            else:
                df_res.at[idx, "is_excluded"] = True

    mask_exc = (df_res["is_excluded"] == True) & (df_res["Вид_отсутствия"].isna() | (df_res["Вид_отсутствия"] == ""))
    df_res.loc[mask_exc, "Вид_отсутствия"] = "Исключение"
    df_res.loc[mask_exc, "причина отсутствия"] = "Исключение"

    return df_res


def calculate_summary_metrics(df_merged: pd.DataFrame) -> Dict[str, Any]:
    total_staff = len(df_merged)

    came_to_office_mask = (df_merged["Начало_дня"].astype(str).str.strip().ne("Нет входа")) & (df_merged.get("is_excluded", False) == False)
    exc_without_doc_mask = (df_merged.get("is_excluded", False) == True) & (
        df_merged["Вид_отсутствия"].isna() | 
        df_merged["Вид_отсутствия"].astype(str).str.strip().isin(["", "nan", "Исключение"])
    )

    working_in_office_count = len(df_merged[came_to_office_mask | exc_without_doc_mask])

    df_not_working = df_merged[~came_to_office_mask & ~exc_without_doc_mask]

    reason_series = df_not_working["причина отсутствия"].astype(str).str.lower()
    is_remote_mask = reason_series.str.contains("удален|дистанцион", regex=True, na=False)
    remote_home = df_not_working[is_remote_mask]
    remote_home_count = len(remote_home)

    df_remaining_absent = df_not_working[~is_remote_mask]
    has_doc_mask = df_remaining_absent["причина отсутствия"].notna() & \
                   df_remaining_absent["причина отсутствия"].ne("") & \
                   df_remaining_absent["причина отсутствия"].ne("nan") & \
                   (~df_remaining_absent["причина отсутствия"].astype(str).str.startswith("Исключение"))
    official_absent_count = len(df_remaining_absent[has_doc_mask])

    unknown = df_remaining_absent[~has_doc_mask]
    unknown_count = len(unknown)

    return {
        "total_staff": total_staff,
        "working_in_office_count": working_in_office_count,
        "remote_home_count": remote_home_count,
        "official_absent_count": official_absent_count,
        "unknown_count": unknown_count,
        "unknown_list": unknown[["fio_clean", "Подразделение", "Должность"]].to_dict(orient="records") if not unknown.empty else []
    }