"""Импорт месячных JSON из report-tool/ в WorkLog."""
from __future__ import annotations

import calendar
import json
import re
from dataclasses import dataclass, field
from datetime import date, datetime
from pathlib import Path

from django.contrib.auth import get_user_model
from django.db import transaction

from apps.it_manager.models import WorkLog
from apps.locations.models.base import Location, organization_object_letter
from apps.locations.models.pharmacies import Pharmacy
from core.models.directories import Directory

Employee = get_user_model()

UA_LOWER_TO_RU = {'і': 'и', 'ї': 'и', 'є': 'е', 'ґ': 'г'}

STAFF_LAST_NAMES = {
    'babich': ('бабич',),
    'aprodov': ('апродов',),
    'khizhniak': ('хижняк', 'хіжняк'),
    'kogutenko': ('когутенко',),
}

DIRECTION_TO_WORK_TYPE = {
    'prro': 'software',
    'terminal': 'installation',
    'print': 'repair',
    'network': 'network',
    'cctv': 'installation',
    'pc': 'repair',
    'install': 'installation',
    'access': 'software',
    'infra': 'maintenance',
}

OBJ_CODE_RE = re.compile(r'^([АAПPТTГGЛL])(\d+)$', re.IGNORECASE)
IMPORT_MARKER = 'import:report-tool:{month_id}'


def normalize_name(value: str) -> str:
    value = (value or '').strip().lower()
    return ''.join(UA_LOWER_TO_RU.get(c, c) for c in value)


def default_report_tool_root() -> Path:
    # apps/it_manager/report_import.py → apms/
    return Path(__file__).resolve().parents[3] / 'report-tool'


def load_company(root: Path) -> dict:
    return json.loads((root / 'company.json').read_text(encoding='utf-8'))


def load_month_json(path: Path) -> dict:
    return json.loads(path.read_text(encoding='utf-8'))


def parse_task_date(value: str, year: int, month: int) -> date | None:
    value = (value or '').strip()
    if not value:
        return None
    for fmt in ('%d.%m.%Y', '%d.%m.%y', '%Y-%m-%d'):
        try:
            return datetime.strptime(value, fmt).date()
        except ValueError:
            continue
    m = re.fullmatch(r'(\d{1,2})/(\d{1,2})/(\d{2,4})', value)
    if m:
        mm, dd, yy = int(m.group(1)), int(m.group(2)), int(m.group(3))
        if yy < 100:
            yy += 2000
        try:
            return date(yy, mm, dd)
        except ValueError:
            return None
    m = re.fullmatch(r'(\d{1,2})\.(\d{1,2})', value)
    if m:
        try:
            return date(year, int(m.group(2)), int(m.group(1)))
        except ValueError:
            return None
    return None


def month_end(year: int, month: int) -> date:
    return date(year, month, calendar.monthrange(year, month)[1])


def classify_work_type_code(text: str, directions: list[dict]) -> str:
    low = (text or '').lower()
    for direction in directions:
        for kw in direction.get('keywords') or []:
            if kw.lower() in low:
                return DIRECTION_TO_WORK_TYPE.get(direction.get('id'), 'other')
    return 'other'


@dataclass
class ImportStats:
    created: int = 0
    skipped: int = 0
    missing_employees: list[str] = field(default_factory=list)
    unresolved_objects: list[str] = field(default_factory=list)
    unmatched_phrases: list[str] = field(default_factory=list)
    months: list[str] = field(default_factory=list)


class CatalogMatchError(Exception):
    """Не все фразы сматчились в справочник — импорт запрещён."""

    def __init__(self, phrases: list[str]):
        self.phrases = phrases
        sample = ', '.join(phrases[:15])
        more = f' … (+{len(phrases) - 15})' if len(phrases) > 15 else ''
        super().__init__(
            f'Нет матча в справочнике для {len(phrases)} фраз: {sample}{more}. '
            f'Дополните work-import-aliases.json и проверьте '
            f'manage.py match_report_tool_works.'
        )


class LocationResolver:
    def __init__(self):
        self._by_prefix_num: dict[tuple[str, int], Location] = {}
        self._by_num: dict[int, list[Location]] = {}
        self._office: Location | None = None
        self._warehouse: Location | None = None
        self._vendors_by_name: dict[str, Location] = {}
        self._built = False

    def build(self):
        if self._built:
            return
        self._built = True
        for pharmacy in Pharmacy.objects.select_related('organization').filter(is_active=True):
            num = pharmacy.pharmacy_number
            if num is None:
                continue
            letter = organization_object_letter(pharmacy.organization)
            if letter:
                self._by_prefix_num[(letter[0].lower(), int(num))] = pharmacy
            self._by_num.setdefault(int(num), []).append(pharmacy)
        self._office = Location.objects.filter(location_type='office', is_active=True).order_by('id').first()
        self._warehouse = Location.objects.filter(location_type='warehouse', is_active=True).order_by('id').first()
        self._vendors_by_name = {}
        for loc in Location.objects.filter(location_type='vendor', is_active=True):
            key = normalize_name(loc.name or '')
            if key:
                self._vendors_by_name[key] = loc

    def resolve(self, object_code: str) -> Location | None:
        self.build()
        raw = (object_code or '').strip()
        if not raw:
            return None
        upper = raw.upper().replace(' ', '')
        if upper in ('ОФИС', 'ОФІС', 'OFFICE'):
            return self._office
        if upper in ('СКЛАД',):
            return self._warehouse
        vendor = self._vendors_by_name.get(normalize_name(raw))
        if vendor is not None:
            return vendor
        m = OBJ_CODE_RE.fullmatch(raw.replace(' ', ''))
        if not m:
            return None
        letter = m.group(1).upper().translate(str.maketrans('APTGL', 'АПТГЛ'))
        num = int(m.group(2))
        letter_l = letter.lower()
        hit = self._by_prefix_num.get((letter_l, num))
        if hit:
            return hit
        # fallback: единственная аптека с таким номером
        candidates = self._by_num.get(num) or []
        if len(candidates) == 1:
            return candidates[0]
        return None


def resolve_employee(staff_id: str, staff_name: str, username_map: dict[str, str]) -> Employee | None:
    if staff_id in username_map:
        return Employee.objects.filter(username=username_map[staff_id]).first()

    last_names = STAFF_LAST_NAMES.get(staff_id)
    if last_names:
        for emp in Employee.objects.filter(is_active=True):
            if normalize_name(emp.last_name) in {normalize_name(n) for n in last_names}:
                return emp

    # по фамилии из полного имени
    parts = (staff_name or '').split()
    if parts:
        target = normalize_name(parts[0])
        for emp in Employee.objects.filter(is_active=True):
            if normalize_name(emp.last_name) == target:
                return emp
    return None


def ensure_employee(staff_id: str, staff_name: str) -> Employee:
    parts = (staff_name or staff_id).split()
    last_name = parts[0] if parts else staff_id
    first_name = parts[1] if len(parts) > 1 else staff_id
    username = f'it.{staff_id}'
    emp, created = Employee.objects.get_or_create(
        username=username,
        defaults={
            'last_name': last_name,
            'first_name': first_name,
            'email': f'{username}@company.local',
            'hire_date': date(2020, 1, 1),
            'user_location_type': 'office',
            'mattermost_sync_enabled': False,
            'ad_sync_enabled': False,
            'is_active': True,
        },
    )
    if created:
        emp.set_unusable_password()
        emp.save()
        from django.contrib.auth.models import Group
        it_group, _ = Group.objects.get_or_create(name='IT')
        emp.groups.add(it_group)
    return emp


def work_type_id(code: str, cache: dict[str, int]) -> int:
    if code not in cache:
        pk = Directory.objects.filter(
            directory_type__code='worklog_type', code=code,
        ).values_list('id', flat=True).first()
        if not pk:
            pk = Directory.objects.filter(
                directory_type__code='worklog_type', code='other',
            ).values_list('id', flat=True).first()
        cache[code] = pk
    return cache[code]


def import_marker(month_id: str) -> str:
    return IMPORT_MARKER.format(month_id=month_id)


def iter_month_files(root: Path, months: list[str] | None) -> list[Path]:
    data_dir = root / 'data'
    files = sorted(data_dir.glob('????-??.json'))
    if months:
        wanted = set(months)
        files = [p for p in files if p.stem in wanted]
    return files


def catalog_work_pk(code: str, cache: dict[str, int]) -> int:
    if code not in cache:
        pk = Directory.objects.filter(
            directory_type__code='it_work', code=code, is_active=True,
        ).exclude(code='_meta').values_list('id', flat=True).first()
        if not pk:
            raise CatalogMatchError([f'(нет Directory it_work code={code})'])
        cache[code] = pk
    return cache[code]


@transaction.atomic
def import_report_tool(
    *,
    root: Path | None = None,
    months: list[str] | None = None,
    username_map: dict[str, str] | None = None,
    create_missing_employees: bool = False,
    dry_run: bool = False,
    require_catalog: bool = True,
    mapping_path: Path | None = None,
    aliases_path: Path | None = None,
) -> ImportStats:
    """Импорт data/YYYY-MM.json → WorkLog с обязательным catalog_work.

    require_catalog=True (по умолчанию): все фрагменты должны сматчиться
    exact по mapping+aliases; иначе CatalogMatchError, записи не пишутся.
    """
    from apps.it_manager.work_import_match import (
        audit_month_files,
        build_phrase_index,
        match_system,
        split_work_text,
    )

    root = root or default_report_tool_root()
    username_map = username_map or {}
    locations = LocationResolver()
    catalog_cache: dict[str, int] = {}
    stats = ImportStats()

    index = build_phrase_index(mapping_path=mapping_path, aliases_path=aliases_path)
    if require_catalog:
        audit = audit_month_files(root / 'data', index, months=months)
        if audit.unmatched:
            phrases = [
                p for p, _ in sorted(
                    audit.unmatched_freq.items(),
                    key=lambda x: (-x[1], x[0].casefold()),
                )
            ]
            stats.unmatched_phrases = phrases
            raise CatalogMatchError(phrases)

    for path in iter_month_files(root, months):
        month = load_month_json(path)
        month_id = month.get('id') or path.stem
        year = int(month['year'])
        month_num = int(month['month'])
        marker = import_marker(month_id)
        stats.months.append(month_id)
        default_date = month_end(year, month_num)

        for emp_data in month.get('employees') or []:
            staff_id = emp_data.get('id') or ''
            staff_name = emp_data.get('name') or staff_id
            employee = resolve_employee(staff_id, staff_name, username_map)
            if employee is None:
                if create_missing_employees:
                    if dry_run:
                        employee = None
                    else:
                        employee = ensure_employee(staff_id, staff_name)
                else:
                    if staff_id not in stats.missing_employees:
                        stats.missing_employees.append(staff_id or staff_name)
                    continue

            for task in emp_data.get('tasks') or []:
                work = (task.get('work') or '').strip()
                if not work:
                    continue
                obj = (task.get('object') or '').strip()
                log_date = parse_task_date(task.get('date') or '', year, month_num) or default_date
                location = locations.resolve(obj) if obj else None
                if obj and location is None and obj.upper() not in stats.unresolved_objects:
                    stats.unresolved_objects.append(obj)

                for part in split_work_text(work):
                    hit = index.match(part)
                    if hit is None:
                        # require_catalog уже отсёк; на всякий случай
                        raise CatalogMatchError([part])
                    _create_or_skip(
                        stats=stats,
                        dry_run=dry_run,
                        employee=employee,
                        log_date=log_date,
                        location=location,
                        catalog_work_id=catalog_work_pk(hit.target_id, catalog_cache),
                        details=hit.details,
                        description=part,
                        marker=marker,
                    )

            for system in emp_data.get('systems') or []:
                name = (system.get('name') or '').strip()
                work = (system.get('work') or '').strip()
                if not name or not work:
                    continue
                hit = match_system(index, name, work)
                if hit is None:
                    raise CatalogMatchError([f'{name}: {work}'])
                description = f'{name}: {work}'
                _create_or_skip(
                    stats=stats,
                    dry_run=dry_run,
                    employee=employee,
                    log_date=default_date,
                    location=locations.resolve('ОФИС'),
                    catalog_work_id=catalog_work_pk(hit.target_id, catalog_cache),
                    details=hit.details,
                    description=description,
                    marker=marker,
                )

    return stats


def _create_or_skip(
    *,
    stats,
    dry_run,
    employee,
    log_date,
    location,
    catalog_work_id,
    details,
    description,
    marker,
):
    if employee is None:
        # dry-run без создания сотрудников
        stats.created += 1
        return
    details = details or {}
    exists = WorkLog.objects.filter(
        technician=employee,
        date=log_date,
        catalog_work_id=catalog_work_id,
        description=description,
        result=marker,
    ).exists()
    if exists:
        stats.skipped += 1
        return
    if dry_run:
        stats.created += 1
        return
    WorkLog.objects.create(
        date=log_date,
        technician=employee,
        location=location,
        catalog_work_id=catalog_work_id,
        details=details,
        work_type_id=None,
        description=description,
        result=marker,
        hours_spent=None,
    )
    stats.created += 1