dak.c2s/backend/app/services/icd_service.py

"""ICD service — normalize, split, validate, save, and generate coding templates."""

import logging
from datetime import datetime, timezone
from io import BytesIO
from typing import Optional

from openpyxl import Workbook, load_workbook
from sqlalchemy.orm import Session

from app.config import get_settings
from app.models.case import Case, CaseICDCode
from app.utils.validators import normalize_icd_hauptgruppe, split_icd_codes, validate_icd

settings = get_settings()

logger = logging.getLogger(__name__)


def normalize_and_validate_icd(raw: str) -> list[tuple[str, str]]:
    """Split, validate, and normalize ICD codes from a raw string.

    Returns list of (icd_code, hauptgruppe) tuples.
    Raises ValueError for any invalid code.
    """
    codes = split_icd_codes(raw)
    result = []
    for code in codes:
        validated = validate_icd(code)
        hauptgruppe = normalize_icd_hauptgruppe(validated)
        result.append((validated, hauptgruppe))
    return result


def save_icd_for_case(
    db: Session,
    case_id: int,
    icd_raw: str,
    user_id: int,
) -> Case:
    """Set ICD codes for a case. Replaces existing ICD codes."""
    case = db.query(Case).filter(Case.id == case_id).first()
    if not case:
        from app.core.exceptions import CaseNotFoundError

        raise CaseNotFoundError()

    # Validate all codes first
    icd_pairs = normalize_and_validate_icd(icd_raw)

    # Delete existing ICD codes for this case
    db.query(CaseICDCode).filter(CaseICDCode.case_id == case_id).delete()

    # Store raw ICD string on case
    case.icd = ", ".join(code for code, _ in icd_pairs)
    case.icd_entered_by = user_id
    case.icd_entered_at = datetime.now(timezone.utc)

    # Create individual ICD code entries
    for code, hauptgruppe in icd_pairs:
        db.add(
            CaseICDCode(
                case_id=case_id,
                icd_code=code,
                icd_hauptgruppe=hauptgruppe,
            )
        )

    db.commit()
    db.refresh(case)
    return case


def get_pending_icd_cases(
    db: Session,
    jahr: Optional[int] = None,
    fallgruppe: Optional[str] = None,
    page: int = 1,
    per_page: int = 50,
) -> tuple[list[Case], int]:
    """Get cases without ICD codes."""
    query = db.query(Case).filter(
        Case.versicherung == settings.VERSICHERUNG_FILTER,
        Case.icd == None,  # noqa: E711
    )

    if jahr:
        query = query.filter(Case.jahr == jahr)
    if fallgruppe:
        query = query.filter(Case.fallgruppe == fallgruppe)

    total = query.count()
    cases = (
        query.order_by(Case.datum.desc())
        .offset((page - 1) * per_page)
        .limit(per_page)
        .all()
    )
    return cases, total


def generate_coding_template(
    db: Session,
    jahr: Optional[int] = None,
    fallgruppe: Optional[str] = None,
) -> bytes:
    """Generate an Excel template for ICD coding.

    Returns .xlsx bytes with columns:
    Case_ID, Fall_ID, Fallgruppe, Datum, ICD (empty)

    Patient names are excluded for data privacy (DSGVO).
    """
    cases, _ = get_pending_icd_cases(
        db, jahr=jahr, fallgruppe=fallgruppe, page=1, per_page=10000
    )

    wb = Workbook()
    ws = wb.active
    ws.title = "ICD Coding"

    # Header
    headers = ["Case_ID", "Fall_ID", "Fallgruppe", "Datum", "ICD"]
    for col, header in enumerate(headers, start=1):
        ws.cell(row=1, column=col, value=header)

    # Data
    for i, case in enumerate(cases, start=2):
        ws.cell(row=i, column=1, value=case.id)
        ws.cell(row=i, column=2, value=case.fall_id)
        ws.cell(row=i, column=3, value=case.fallgruppe)
        ws.cell(row=i, column=4, value=case.datum.isoformat() if case.datum else "")
        # Column 5 (ICD) left empty for admin to fill in

    # Auto-width
    for col in ws.columns:
        max_length = max(len(str(cell.value or "")) for cell in col)
        ws.column_dimensions[col[0].column_letter].width = min(max_length + 2, 30)

    buffer = BytesIO()
    wb.save(buffer)
    return buffer.getvalue()


def import_icd_from_xlsx(db: Session, content: bytes, user_id: int) -> dict:
    """Import ICD codes from an Excel file.

    Auto-detects the format:
      - Coding template: Column A contains numeric Case_IDs
      - Wochenübersicht: Column A contains KVNR strings (letter prefix)

    Returns: {"updated": int, "errors": list[str]}
    """
    wb = load_workbook(BytesIO(content), read_only=True)
    ws = wb.active

    # Detect format by scanning for first data row
    detected_format = _detect_import_format(ws)

    if detected_format == "wochenuebersicht":
        return _import_icd_wochenuebersicht(db, ws, user_id)

    # Default: Coding template (Case_ID-based)
    return _import_icd_coding_template(db, ws, user_id)


def _detect_import_format(ws) -> str:
    """Detect whether the worksheet is a coding template or Wochenübersicht."""
    for row in ws.iter_rows(min_row=1, max_row=20, values_only=True):
        if not row or not row[0]:
            continue
        val = str(row[0]).strip()
        # Skip header rows
        if val.lower() in ("case_id", "kvnr", "kalenderwoche:"):
            continue
        # If it's a pure integer, it's a coding template (Case_ID)
        try:
            int(val)
            return "coding_template"
        except (ValueError, TypeError):
            pass
        # If it starts with a letter, it's likely a KVNR → Wochenübersicht
        if val and val[0].isalpha():
            return "wochenuebersicht"
    return "coding_template"


def _import_icd_coding_template(db: Session, ws, user_id: int) -> dict:
    """Import ICD codes from a coding template (Case_ID in column A)."""
    updated = 0
    errors: list[str] = []

    for row in ws.iter_rows(min_row=2, values_only=False):
        case_id_cell = row[0].value
        if not case_id_cell:
            continue

        try:
            case_id = int(case_id_cell)
        except (ValueError, TypeError):
            continue

        # Find ICD column: last column (col 5 in new template, col 7 in legacy)
        icd_value = None
        last_idx = len(row) - 1
        if last_idx >= 0 and row[last_idx].value:
            icd_value = str(row[last_idx].value).strip()

        if not icd_value:
            continue

        try:
            save_icd_for_case(db, case_id, icd_value, user_id)
            updated += 1
        except Exception as e:
            errors.append(f"Case {case_id}: {e}")

    return {"updated": updated, "errors": errors}


def _import_icd_wochenuebersicht(db: Session, ws, user_id: int) -> dict:
    """Import ICD codes from a Wochenübersicht file (KVNR in col A, ICD in col I)."""
    updated = 0
    errors: list[str] = []

    for row in ws.iter_rows(values_only=False):
        # Skip rows without enough columns
        if len(row) < 9:
            continue

        col_a = row[0].value
        if not col_a:
            continue

        col_a_str = str(col_a).strip()

        # Skip header/label rows
        if col_a_str.lower() in ("kvnr", ""):
            continue
        # Skip non-data rows (e.g. "Kalenderwoche:", instruction text)
        if not col_a_str[0].isalpha() and not col_a_str[0].isdigit():
            continue
        # Skip rows where col A looks like a header label
        if col_a_str.lower().startswith(("bitte", "kalender")):
            continue

        kvnr = col_a_str

        # ICD is in column I (index 8)
        icd_cell = row[8].value
        if not icd_cell:
            continue
        icd_value = str(icd_cell).strip()
        if not icd_value or icd_value.lower() == "icd-10":
            continue

        # Find case by KVNR — take the most recent one
        case = (
            db.query(Case)
            .filter(
                Case.kvnr == kvnr,
                Case.versicherung == settings.VERSICHERUNG_FILTER,
            )
            .order_by(Case.datum.desc())
            .first()
        )

        if not case:
            errors.append(f"KVNR {kvnr}: Kein Fall gefunden")
            continue

        try:
            save_icd_for_case(db, case.id, icd_value, user_id)
            updated += 1
        except Exception as e:
            errors.append(f"KVNR {kvnr} (Case {case.id}): {e}")

    return {"updated": updated, "errors": errors}