"""Historical Excel import for Abrechnung_DAK.xlsx.

Imports case data from the master Excel workbook into the database.
Each year-sheet (2026, 2025, 2024, 2023, 2020-2022) is imported independently.
Sheets like 'Gutachten', 'Ubersicht', 'BKK Salzgitter', '_2023', and 'Tabelle1'
are skipped.

Column mapping is dynamic -- headers are read from row 1 and matched by name,
so column order changes between sheets are handled automatically. The '2020-2022'
sheet has an extra 'Jahr' column; for single-year sheets the year is derived
from the sheet name.

Fallgruppe normalization maps the messy Excel values to the 5 valid DB codes:
onko, kardio, intensiv, galle, sd.
"""

from __future__ import annotations

import datetime as dt
import logging
import os
from typing import Any

from openpyxl import load_workbook
from sqlalchemy.orm import Session

from app.models.audit import ImportLog
from app.models.case import Case

logger = logging.getLogger(__name__)

# ---------------------------------------------------------------------------
# Fallgruppe normalization
# ---------------------------------------------------------------------------
# The Excel contains a messy mix of values that must map to the 5 valid DB
# codes enforced by CHECK constraint: onko, kardio, intensiv, galle, sd

_FALLGRUPPE_MAP: dict[str, str] = {
    # Standard codes
    "onko": "onko",
    "kardio": "kardio",
    "intensiv": "intensiv",
    "galle": "galle",
    "sd": "sd",
    "schild": "sd",  # Schilddruese -> sd
    # Typos observed in data
    "intei": "intensiv",
    "intsiv": "intensiv",
    # Non-standard categories mapped to closest valid code
    "medi": "onko",  # Medizin-Onko cases
    "radio": "onko",  # Radiologie-Onko cases
    "knie": "intensiv",  # Orthopedic, mapped to intensiv
    "schmerz": "intensiv",  # Pain cases, mapped to intensiv
    "wunde": "intensiv",  # Wound cases, mapped to intensiv
    "orthopaedie": "intensiv",
    "orthopadie": "intensiv",
}

# Keyword-based fallback for compound values like "onko brust", "onko Lymph"
_FALLGRUPPE_KEYWORDS: list[tuple[str, str]] = [
    ("onko", "onko"),
    ("kardio", "kardio"),
    ("intensiv", "intensiv"),
    ("galle", "galle"),
    ("schild", "sd"),
    ("sd", "sd"),
]


def _normalize_fallgruppe(raw: str | None) -> str | None:
    """Map a raw Fallgruppe string to a valid DB code.

    Returns None if the value cannot be mapped (caller decides how to handle).
    """
    if not raw:
        return None

    cleaned = raw.strip().lower()
    # Remove accents for orthopaedie matching
    cleaned_ascii = cleaned.replace("\u00e4", "ae").replace("\u00f6", "oe").replace("\u00fc", "ue")

    # Direct match
    if cleaned in _FALLGRUPPE_MAP:
        return _FALLGRUPPE_MAP[cleaned]
    if cleaned_ascii in _FALLGRUPPE_MAP:
        return _FALLGRUPPE_MAP[cleaned_ascii]

    # Keyword match (for compound values like "onko brust")
    for keyword, code in _FALLGRUPPE_KEYWORDS:
        if keyword in cleaned:
            return code

    return None


# ---------------------------------------------------------------------------
# Boolean parsing
# ---------------------------------------------------------------------------

def _parse_bool(value: Any) -> bool:
    """Parse a cell value to boolean.

    Handles: True, False, None, 'Ja', 'Nein', 'nix', ' ', 1, 0, etc.
    """
    if value is None:
        return False
    if isinstance(value, bool):
        return value
    if isinstance(value, (int, float)):
        return bool(value)
    if isinstance(value, str):
        v = value.strip().lower()
        if v in ("ja", "yes", "1", "true", "x"):
            return True
        # Everything else (nein, no, nix, empty, space) -> False
        return False
    return False


# ---------------------------------------------------------------------------
# Date parsing
# ---------------------------------------------------------------------------

def _parse_date(value: Any) -> dt.date | None:
    """Parse a cell value to a date.

    openpyxl typically returns datetime objects for date-formatted cells.
    Strings in DD.MM.YYYY format are also handled.
    """
    if value is None:
        return None
    if isinstance(value, dt.datetime):
        return value.date()
    if isinstance(value, dt.date):
        return value
    if isinstance(value, str):
        value = value.strip()
        if not value:
            return None
        # Try DD.MM.YYYY
        parts = value.split(".")
        if len(parts) == 3:
            try:
                day, month, year = int(parts[0]), int(parts[1]), int(parts[2])
                if year < 100:
                    year = 2000 + year if year < 50 else 1900 + year
                return dt.date(year, month, day)
            except (ValueError, TypeError):
                pass
        logger.warning("Cannot parse date value: %r", value)
        return None
    if isinstance(value, (int, float)):
        # Excel serial date number -- openpyxl normally converts these,
        # but just in case, handle it.
        try:
            # Excel epoch is 1899-12-30 (with the Lotus 1-2-3 bug)
            return (dt.datetime(1899, 12, 30) + dt.timedelta(days=int(value))).date()
        except (ValueError, OverflowError):
            logger.warning("Cannot parse date from number: %r", value)
            return None
    return None


# ---------------------------------------------------------------------------
# String helper
# ---------------------------------------------------------------------------

def _str_or_none(value: Any, max_len: int | None = None) -> str | None:
    """Convert cell value to stripped string or None."""
    if value is None:
        return None
    s = str(value).strip()
    if not s:
        return None
    if max_len and len(s) > max_len:
        s = s[:max_len]
    return s


def _phone_str(value: Any) -> str | None:
    """Convert phone number cell to string.

    Phone numbers in Excel are often stored as integers (e.g. 4915121659287).
    We need to convert them to strings, optionally adding a '+' prefix.
    """
    if value is None:
        return None
    if isinstance(value, (int, float)):
        s = str(int(value))
        # If it starts with 49 and is long enough, add + prefix
        if len(s) >= 10 and s.startswith("49"):
            return "+" + s
        return s
    s = str(value).strip()
    return s if s else None


# ---------------------------------------------------------------------------
# Fall-ID generation (matches import_service.generate_fall_id format)
# ---------------------------------------------------------------------------

def _generate_fall_id(jahr: int, kw: int, fallgruppe: str, nachname: str) -> str:
    """Generate fall_id: YYYY-KW02d-fallgruppe-Nachname."""
    return f"{jahr}-{kw:02d}-{fallgruppe}-{nachname}"


# ---------------------------------------------------------------------------
# Sheet import
# ---------------------------------------------------------------------------

# Standard header names (case-insensitive matching)
_HEADER_MAP = {
    "id": "id",
    "jahr": "jahr",
    "kw": "kw",
    "datum": "datum",
    "anrede": "anrede",
    "vorname": "vorname",
    "nachname": "nachname",
    "geburtsdatum": "geburtsdatum",
    "kvnr": "kvnr",
    "versicherung": "versicherung",
    "icd": "icd",
    "fallgruppe": "fallgruppe",
    "strasse": "strasse",
    "strasze": "strasse",
    "stra\u00dfe": "strasse",  # Strasse with eszett
    "plz": "plz",
    "ort": "ort",
    "e-mail": "email",
    "email": "email",
    "ansprechpartner": "ansprechpartner",
    "telefonnummer": "telefonnummer",
    "mobiltelefon": "mobiltelefon",
    "unterlagen": "unterlagen",
    "unterlagen verschickt": "unterlagen_verschickt",
    "erhalten": "erhalten",
    "unterlagen erhalten": "unterlagen_erhalten",
    "unterlagen an gutachter": "unterlagen_an_gutachter",
    "gutachten": "gutachten",
    "gutachter": "gutachter",
    "gutachten erstellt": "gutachten_erstellt",
    "gutachten versendet": "gutachten_versendet",
    "schweigepflicht": "schweigepflicht",
    "ablehnung": "ablehnung",
    "abbruch": "abbruch",
    "abbruch_datum": "abbruch_datum",
    "kurzbeschreibung": "kurzbeschreibung",
    "fragestellung": "fragestellung",
    "kommentar": "kommentar",
    "e-mail2": "email2",
    "email2": "email2",
    "telefon2": "telefon2",
    "sonstiges": "sonstiges",
    "abgerechnet": "abgerechnet",
    "abrechnung_datum": "abrechnung_datum",
}

# Sheets to import (in order)
YEAR_SHEETS = ["2020-2022", "2023", "2024", "2025", "2026"]

# Sheets to skip
SKIP_SHEETS = {"Gutachten", "\u00dcbersicht", "Ubersicht", "BKK Salzgitter",
               "_2023", "Tabelle1"}


def _build_col_map(header_row: tuple) -> dict[str, int]:
    """Build mapping from canonical field name -> column index (0-based).

    Reads header row and matches each cell against _HEADER_MAP.
    """
    col_map: dict[str, int] = {}
    for idx, cell_value in enumerate(header_row):
        if cell_value is None:
            continue
        key = str(cell_value).strip().lower()
        canonical = _HEADER_MAP.get(key)
        if canonical and canonical not in col_map:
            col_map[canonical] = idx
    return col_map


def _get(row: tuple, col_map: dict[str, int], field: str) -> Any:
    """Get a value from a row by canonical field name."""
    idx = col_map.get(field)
    if idx is None:
        return None
    if idx >= len(row):
        return None
    return row[idx]


def import_abrechnung_sheet(
    db: Session,
    ws,  # openpyxl worksheet (read-only)
    sheet_name: str,
    default_year: int | None = None,
    user_id: int | None = None,
) -> dict:
    """Import a single sheet from Abrechnung_DAK.xlsx.

    Args:
        db: SQLAlchemy session.
        ws: openpyxl worksheet object.
        sheet_name: Name of the sheet (for logging and import_source).
        default_year: Year to use if not available per-row (derived from sheet name).
        user_id: User ID for import logging.

    Returns:
        {"imported": int, "skipped": int, "errors": list[str]}
    """
    imported = 0
    skipped = 0
    errors: list[str] = []

    # Read header row
    rows_iter = ws.iter_rows(values_only=True)
    try:
        header_row = next(rows_iter)
    except StopIteration:
        return {"imported": 0, "skipped": 0, "errors": ["Empty sheet"]}

    col_map = _build_col_map(header_row)

    # Verify essential columns exist
    required = {"nachname", "fallgruppe", "datum"}
    missing = required - col_map.keys()
    if missing:
        return {
            "imported": 0,
            "skipped": 0,
            "errors": [f"Missing required columns: {missing}"],
        }

    has_jahr_col = "jahr" in col_map

    for row_num, row in enumerate(rows_iter, start=2):
        try:
            nachname_raw = _get(row, col_map, "nachname")
            if not nachname_raw or (isinstance(nachname_raw, str) and not nachname_raw.strip()):
                skipped += 1
                continue

            nachname = str(nachname_raw).strip()

            # Parse datum
            datum = _parse_date(_get(row, col_map, "datum"))
            if datum is None:
                errors.append(f"Row {row_num}: Missing/invalid Datum for {nachname}")
                skipped += 1
                continue

            # Determine year
            if has_jahr_col:
                jahr_val = _get(row, col_map, "jahr")
                if jahr_val is not None:
                    jahr = int(jahr_val)
                else:
                    jahr = default_year or datum.year
            else:
                jahr = default_year or datum.year

            # Parse KW
            kw_val = _get(row, col_map, "kw")
            if kw_val is not None:
                try:
                    kw = int(kw_val)
                except (ValueError, TypeError):
                    kw = datum.isocalendar()[1]
            else:
                kw = datum.isocalendar()[1]

            # Normalize Fallgruppe
            fallgruppe_raw = _str_or_none(_get(row, col_map, "fallgruppe"))
            fallgruppe = _normalize_fallgruppe(fallgruppe_raw)
            if fallgruppe is None:
                errors.append(
                    f"Row {row_num}: Cannot map Fallgruppe '{fallgruppe_raw}' "
                    f"for {nachname}"
                )
                skipped += 1
                continue

            # Generate fall_id
            fall_id = _generate_fall_id(jahr, kw, fallgruppe, nachname)

            # Check for duplicate by fall_id
            existing = db.query(Case.id).filter(Case.fall_id == fall_id).first()
            if existing:
                skipped += 1
                continue

            # Parse all other fields
            vorname = _str_or_none(_get(row, col_map, "vorname"), max_len=100)
            geburtsdatum = _parse_date(_get(row, col_map, "geburtsdatum"))

            # Also check for duplicate by personal data
            dup_query = db.query(Case.id).filter(
                Case.nachname == nachname,
                Case.fallgruppe == fallgruppe,
                Case.datum == datum,
            )
            if vorname:
                dup_query = dup_query.filter(Case.vorname == vorname)
            if geburtsdatum:
                dup_query = dup_query.filter(Case.geburtsdatum == geburtsdatum)
            if dup_query.first():
                skipped += 1
                continue

            anrede = _str_or_none(_get(row, col_map, "anrede"), max_len=20)
            kvnr = _str_or_none(_get(row, col_map, "kvnr"), max_len=20)
            versicherung = _str_or_none(_get(row, col_map, "versicherung"), max_len=50) or "DAK"
            icd = _str_or_none(_get(row, col_map, "icd"))
            strasse = _str_or_none(_get(row, col_map, "strasse"), max_len=255)
            plz_raw = _get(row, col_map, "plz")
            plz = str(int(plz_raw)).zfill(5) if isinstance(plz_raw, (int, float)) else _str_or_none(plz_raw, max_len=10)
            ort = _str_or_none(_get(row, col_map, "ort"), max_len=100)
            email = _str_or_none(_get(row, col_map, "email"), max_len=255)
            ansprechpartner = _str_or_none(_get(row, col_map, "ansprechpartner"), max_len=200)
            telefonnummer = _phone_str(_get(row, col_map, "telefonnummer"))
            if telefonnummer and len(telefonnummer) > 50:
                telefonnummer = telefonnummer[:50]
            mobiltelefon = _phone_str(_get(row, col_map, "mobiltelefon"))
            if mobiltelefon and len(mobiltelefon) > 50:
                mobiltelefon = mobiltelefon[:50]
            email2 = _str_or_none(_get(row, col_map, "email2"), max_len=255)
            telefon2 = _phone_str(_get(row, col_map, "telefon2"))
            if telefon2 and len(telefon2) > 50:
                telefon2 = telefon2[:50]

            unterlagen = _parse_bool(_get(row, col_map, "unterlagen"))
            unterlagen_verschickt = _parse_date(_get(row, col_map, "unterlagen_verschickt"))
            # "erhalten" is Optional[bool] -- None means unknown
            erhalten_raw = _get(row, col_map, "erhalten")
            erhalten = None if erhalten_raw is None else _parse_bool(erhalten_raw)
            unterlagen_erhalten = _parse_date(_get(row, col_map, "unterlagen_erhalten"))
            unterlagen_an_gutachter = _parse_date(_get(row, col_map, "unterlagen_an_gutachter"))
            gutachten_bool = _parse_bool(_get(row, col_map, "gutachten"))
            gutachter = _str_or_none(_get(row, col_map, "gutachter"), max_len=100)
            gutachten_erstellt = _parse_date(_get(row, col_map, "gutachten_erstellt"))
            gutachten_versendet = _parse_date(_get(row, col_map, "gutachten_versendet"))
            schweigepflicht = _parse_bool(_get(row, col_map, "schweigepflicht"))
            ablehnung = _parse_bool(_get(row, col_map, "ablehnung"))
            abbruch = _parse_bool(_get(row, col_map, "abbruch"))
            abbruch_datum = _parse_date(_get(row, col_map, "abbruch_datum"))
            kurzbeschreibung = _str_or_none(_get(row, col_map, "kurzbeschreibung"))
            fragestellung = _str_or_none(_get(row, col_map, "fragestellung"))
            kommentar = _str_or_none(_get(row, col_map, "kommentar"))
            sonstiges = _str_or_none(_get(row, col_map, "sonstiges"))
            abgerechnet = _parse_bool(_get(row, col_map, "abgerechnet"))
            abrechnung_datum = _parse_date(_get(row, col_map, "abrechnung_datum"))

            case = Case(
                fall_id=fall_id,
                jahr=jahr,
                kw=kw,
                datum=datum,
                anrede=anrede,
                vorname=vorname,
                nachname=nachname,
                geburtsdatum=geburtsdatum,
                kvnr=kvnr,
                versicherung=versicherung,
                icd=icd,
                fallgruppe=fallgruppe,
                strasse=strasse,
                plz=plz,
                ort=ort,
                email=email,
                ansprechpartner=ansprechpartner,
                telefonnummer=telefonnummer,
                mobiltelefon=mobiltelefon,
                email2=email2,
                telefon2=telefon2,
                unterlagen=unterlagen,
                unterlagen_verschickt=unterlagen_verschickt,
                erhalten=erhalten,
                unterlagen_erhalten=unterlagen_erhalten,
                unterlagen_an_gutachter=unterlagen_an_gutachter,
                gutachten=gutachten_bool,
                gutachter=gutachter,
                gutachten_erstellt=gutachten_erstellt,
                gutachten_versendet=gutachten_versendet,
                schweigepflicht=schweigepflicht,
                ablehnung=ablehnung,
                abbruch=abbruch,
                abbruch_datum=abbruch_datum,
                kurzbeschreibung=kurzbeschreibung,
                fragestellung=fragestellung,
                kommentar=kommentar,
                sonstiges=sonstiges,
                abgerechnet=abgerechnet,
                abrechnung_datum=abrechnung_datum,
                import_source=f"Abrechnung_DAK.xlsx:{sheet_name}",
            )
            db.add(case)
            imported += 1

            # Flush in batches of 100 to catch constraint violations early
            if imported % 100 == 0:
                db.flush()

        except Exception as e:
            nachname_display = _str_or_none(_get(row, col_map, "nachname")) or "?"
            errors.append(f"Row {row_num} ({nachname_display}): {e}")
            logger.warning(
                "Import error in sheet '%s' row %d: %s",
                sheet_name, row_num, e,
            )

    # Final flush
    if imported > 0:
        db.flush()

    logger.info(
        "Sheet '%s': %d imported, %d skipped, %d errors",
        sheet_name, imported, skipped, len(errors),
    )

    return {"imported": imported, "skipped": skipped, "errors": errors}


def import_full_abrechnung(
    db: Session,
    filepath: str,
    user_id: int | None = None,
) -> dict:
    """Import all relevant sheets from Abrechnung_DAK.xlsx.

    Opens the workbook in read-only mode, iterates through year sheets,
    commits all changes at the end, and logs the import.

    Args:
        db: SQLAlchemy session.
        filepath: Path to the Excel file.
        user_id: User ID for import logging.

    Returns:
        Dict mapping sheet name -> {"imported": int, "skipped": int, "errors": list}
    """
    filename = os.path.basename(filepath)
    wb = load_workbook(filepath, read_only=True, data_only=True)
    results: dict[str, dict] = {}
    total_imported = 0
    total_skipped = 0
    all_errors: list[str] = []

    try:
        available_sheets = set(wb.sheetnames)

        for sheet_name in YEAR_SHEETS:
            if sheet_name not in available_sheets:
                logger.info("Sheet '%s' not found, skipping", sheet_name)
                continue

            # Determine default year from sheet name
            if sheet_name == "2020-2022":
                default_year = None  # Will use per-row "Jahr" column
            else:
                try:
                    default_year = int(sheet_name)
                except ValueError:
                    default_year = None

            ws = wb[sheet_name]
            logger.info("Importing sheet '%s'...", sheet_name)

            result = import_abrechnung_sheet(
                db=db,
                ws=ws,
                sheet_name=sheet_name,
                default_year=default_year,
                user_id=user_id,
            )
            results[sheet_name] = result
            total_imported += result["imported"]
            total_skipped += result["skipped"]
            if result["errors"]:
                all_errors.extend(
                    [f"[{sheet_name}] {e}" for e in result["errors"]]
                )

        # Also check for any additional year-like sheets not in our list
        for sn in wb.sheetnames:
            if sn in SKIP_SHEETS or sn in results:
                continue
            # Check if it looks like a year sheet (has standard headers)
            ws = wb[sn]
            try:
                header = next(ws.iter_rows(max_row=1, values_only=True))
                header_names = {str(h).strip().lower() for h in header if h}
                if "nachname" in header_names and "fallgruppe" in header_names:
                    logger.info(
                        "Found additional data sheet '%s', skipping "
                        "(not in YEAR_SHEETS list). Add it manually if needed.",
                        sn,
                    )
            except StopIteration:
                pass

        # Commit everything
        db.flush()

        # Log the import
        log = ImportLog(
            filename=filename,
            import_type="historical_excel",
            cases_imported=total_imported,
            cases_skipped=total_skipped,
            cases_updated=0,
            errors="; ".join(all_errors[:50]) if all_errors else None,
            details={
                "sheets": {
                    sn: {"imported": r["imported"], "skipped": r["skipped"],
                         "error_count": len(r["errors"])}
                    for sn, r in results.items()
                }
            },
            imported_by=user_id,
        )
        db.add(log)
        db.commit()

    except Exception:
        db.rollback()
        raise
    finally:
        wb.close()

    logger.info(
        "Full import complete: %d imported, %d skipped, %d errors across %d sheets",
        total_imported, total_skipped, len(all_errors), len(results),
    )

    return results