backend/api/routes/compat.py

"""Legacy frontend API compatibility layer backed by v2 services."""

from __future__ import annotations

import logging
import shutil
import tempfile
import time
from pathlib import Path
from typing import Any

from fastapi import (
    APIRouter,
    BackgroundTasks,
    Depends,
    File,
    Header,
    HTTPException,
    Query,
    Response,
    UploadFile,
    status,
)
from starlette.concurrency import run_in_threadpool
from pydantic import BaseModel, Field

from backend.api.deps import get_current_tenant
from backend.api import security
from backend.config import settings
from backend.core import template_discoverer
from backend.core.content_similarity import scan_similar_content
from backend.core import document_library
from backend.core.legacy_generation import run_legacy_generation
from backend.core.notes_extract import extract_notes_from_upload
from backend.core.notes_routing import route_lines_to_l3_sections
from backend.core import photo_store
from backend.core.photo_store import ALLOWED_CONTENT_TYPES
from backend.core.rag_store import TIER_REFERENCE, get_rag_store
from backend.core.report_assembler import to_docx
from backend.core.report_session import (
    ReportSession,
    create_report_session,
    get_document,
    list_documents,
    load_session,
    save_session,
)
logger = logging.getLogger(__name__)

router = APIRouter(tags=["legacy-compat"])

_GROUP_LABELS = {
    "A": "A — About the inspection",
    "B": "B — Overall opinion and summary of ratings",
    "C": "C — About the property",
    "D": "D — Outside the property",
    "E": "E — Inside the property",
    "F": "F — Services",
    "G": "G — Grounds (including shared areas for flats)",
    "H": "H — Issues for your legal advisers",
    "I": "I — Risks",
    "J": "J — Energy matters",
    "K": "K — Surveyor's declaration",
    "L": "L — What to do now",
    "M": "M — Description of the RICS Home Survey - Level 3 service and terms of engagement",
    "N": "N — Typical house diagram",
}


class LegacyGenerateBody(BaseModel):
    template_id: str
    template_ids: list[str] = Field(default_factory=list)
    bullets: list[str] = Field(default_factory=list)
    bullets_by_section: dict[str, list[str]] = Field(default_factory=dict)
    mode: str = "generate"
    interference_level: str | None = None
    retrieval_level: str = "paragraph"
    force_regenerate: bool = True
    strict_uploaded_only: bool = False
    reference_document_ids: list[str] = Field(default_factory=list)
    draft_paragraph: str | None = None


class BatchStatusBody(BaseModel):
    document_ids: list[str] = Field(default_factory=list)


class RouteNotesBody(BaseModel):
    lines: list[str] = Field(default_factory=list, description="Raw note lines to route into L3 sections.")


class SurveyClassifyBody(BaseModel):
    document_ids: list[str] = Field(default_factory=list)
    force_refresh_corpus: bool = False


class SurveyApplyBody(BaseModel):
    items: list[dict[str, Any]] = Field(default_factory=list)


class PhotoAiSelectionBody(BaseModel):
    photo_ids: list[str] = Field(default_factory=list)


class SimilarContentBody(BaseModel):
    text: str = ""
    section_code: str | None = None
    peer_sections: dict[str, str] = Field(default_factory=dict)
    limit: int = 8
    min_relevance_percent: float = 28.0
    exclude_document_ids: list[str] = Field(default_factory=list)


class SectionTextPatch(BaseModel):
    text: str


def _section_group(code: str) -> str:
    code = (code or "").strip().upper()
    if len(code) == 1:
        return code
    return code[0] if code else "A"


def _resolve_tenant_token(
    authorization: str = Header(default=""),
    access_token: str = Query(default=""),
) -> str:
    token = ""
    if authorization.lower().startswith("bearer "):
        token = authorization.split(" ", 1)[1].strip()
    elif access_token.strip():
        token = access_token.strip()
    if not token:
        raise HTTPException(status_code=401, detail="Missing bearer token")
    payload = security.decode_token(token)
    tenant_id = payload.get("sub")
    if not tenant_id:
        raise HTTPException(status_code=401, detail="Token missing subject")
    return tenant_id


def _require_session(tenant_id: str, report_id: str) -> ReportSession:
    session = load_session(tenant_id, report_id)
    if session is None:
        raise HTTPException(status_code=404, detail="Report not found")
    return session


def _legacy_photo_url(report_id: str, section_id: str, photo_id: str) -> str:
    return f"/reports/{report_id}/sections/{section_id}/photos/{photo_id}"


def _photo_rows(tenant_id: str, session: ReportSession, section_id: str) -> list[dict]:
    rows = photo_store.list_section_photos(tenant_id, session.draft_id, section_id)
    return [
        {
            "photo_id": p.id,
            "original_filename": p.filename,
            "url": _legacy_photo_url(session.report_id, section_id, p.id),
            "selected_for_ai": p.selected_for_ai,
        }
        for p in rows
    ]


async def _run_generation(tenant_id: str, report_id: str, body: LegacyGenerateBody) -> None:
    """Run blocking RAG + LLM generation off the async event loop."""
    await run_in_threadpool(run_legacy_generation, tenant_id, report_id, body)


def _ingest_upload_file(
    tenant_id: str,
    tmp_path: Path,
    original_filename: str,
):
    return document_library.ingest_and_register(
        tenant_id,
        tmp_path,
        original_filename=original_filename,
    )


# ── Upload & documents ───────────────────────────────────────────────────────


@router.post("/upload/batch", status_code=status.HTTP_201_CREATED)
async def upload_batch(
    files: list[UploadFile] = File(...),
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    from backend.ingest.zip_extract import extract_reference_documents

    allowed_ext = {".pdf", ".docx", ".docm", ".doc", ".zip"}
    items: list[dict] = []
    accepted = 0
    rejected = 0

    for upload in files:
        suffix = Path(upload.filename or "").suffix.lower()
        if suffix not in allowed_ext:
            rejected += 1
            items.append({
                "document_id": None,
                "filename": upload.filename,
                "status": "rejected",
                "message": f"Unsupported type {suffix}. Use PDF, Word, or ZIP.",
            })
            continue

        upload_tmp_root = settings.data_dir_path / "tmp" / "uploads"
        from backend.utils.runtime_paths import ensure_data_drive_runtime_dirs

        ensure_data_drive_runtime_dirs()
        upload_tmp_root.mkdir(parents=True, exist_ok=True)
        tmp_dir = Path(tempfile.mkdtemp(prefix="v2_batch_", dir=str(upload_tmp_root)))
        safe_name = Path(upload.filename or f"upload{suffix}").name
        tmp_path = tmp_dir / safe_name
        try:
            content = await upload.read()
            if not content:
                raise ValueError("Empty upload body")
            tmp_path.write_bytes(content)

            if suffix == ".zip":
                extract_dir = tmp_dir / "extracted"
                extracted = extract_reference_documents(tmp_path, extract_dir)
                if not extracted:
                    rejected += 1
                    items.append({
                        "document_id": None,
                        "filename": upload.filename,
                        "status": "rejected",
                        "message": "ZIP contained no supported PDF or Word files.",
                    })
                    continue
                for _inner, extracted_path in extracted:
                    doc = _ingest_upload_file(
                        tenant_id,
                        extracted_path,
                        extracted_path.name,
                    )
                    accepted += 1
                    items.append({
                        "document_id": doc.document_id,
                        "filename": doc.filename,
                        "status": "complete",
                        "message": f"Ingested {doc.ingested_chunks} reference chunks.",
                    })
            else:
                if tmp_path.stat().st_size > settings.max_single_upload_bytes:
                    rejected += 1
                    items.append({
                        "document_id": None,
                        "filename": upload.filename,
                        "status": "rejected",
                        "message": f"File exceeds max size ({settings.max_single_upload_bytes} bytes).",
                    })
                    continue
                doc = _ingest_upload_file(
                    tenant_id,
                    tmp_path,
                    upload.filename or tmp_path.name,
                )
                accepted += 1
                items.append({
                    "document_id": doc.document_id,
                    "filename": upload.filename,
                    "status": "complete",
                    "message": f"Ingested {doc.ingested_chunks} reference chunks.",
                })
        except Exception as exc:  # noqa: BLE001
            logger.exception("Batch upload failed for %s", upload.filename)
            rejected += 1
            items.append({
                "document_id": None,
                "filename": upload.filename,
                "status": "failed",
                "message": str(exc),
            })
        finally:
            shutil.rmtree(tmp_dir, ignore_errors=True)

    return {
        "accepted": accepted,
        "rejected": rejected,
        "message": f"{accepted} file(s) indexed as past-report reference.",
        "items": items,
    }


@router.post("/documents/batch-status")
def batch_status(
    body: BatchStatusBody,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    docs = list_documents(tenant_id)
    items = []
    complete = failed = pending = processing = 0
    for doc_id in body.document_ids:
        row = docs.get(doc_id)
        if row is None:
            failed += 1
            items.append({
                "document_id": doc_id,
                "status": "not_found",
                "filename": "",
                "error": "Unknown document id",
            })
            continue
        st = row.status
        if st == "complete":
            complete += 1
        elif st == "failed":
            failed += 1
        else:
            pending += 1
        items.append({
            "document_id": doc_id,
            "status": st,
            "filename": row.filename,
            "error": row.error,
        })
    return {
        "pending": pending,
        "processing": processing,
        "complete": complete,
        "failed": failed,
        "items": items,
    }


@router.get("/documents/tenant-chunk-summary")
def tenant_chunk_summary(tenant_id: str = Depends(get_current_tenant)) -> dict:
    count = get_rag_store().count(tenant_id, TIER_REFERENCE)
    return {"tenant_id": tenant_id, "indexed_chunk_count": count}


@router.post("/documents/survey-level/classify")
def survey_level_classify(
    body: SurveyClassifyBody,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    docs = list_documents(tenant_id)
    items = []
    for doc_id in body.document_ids:
        row = docs.get(doc_id)
        items.append({
            "document_id": doc_id,
            "filename": row.filename if row else doc_id,
            "predicted_survey_level": 3,
            "confidence": 0.85,
            "rationale": "Inferred from template schema (v2 reference mapping).",
        })
    return {"items": items, "corpus_labelled_files": len(items)}


@router.post("/documents/survey-level/apply")
def survey_level_apply(
    body: SurveyApplyBody,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    return {"updated": body.items}


# ── Reports ──────────────────────────────────────────────────────────────────


@router.post("/reports")
def create_report(
    document_id: str,
    survey_level: int = 3,
    confirm_tier_mismatch: bool = False,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    if not get_document(tenant_id, document_id):
        raise HTTPException(status_code=404, detail="Document not found")
    template_discoverer.ensure_canonical_schema(tenant_id)
    if get_rag_store().count(tenant_id, TIER_REFERENCE) == 0:
        raise HTTPException(
            status_code=400,
            detail="Upload at least one past report before creating a report job.",
        )
    session = create_report_session(
        tenant_id,
        survey_level=max(1, min(3, survey_level)),
        primary_document_id=document_id,
        document_ids=[document_id],
    )
    return {"report_id": session.report_id, "status": "ready", "message": "Report session created."}


@router.get("/templates/catalog")
def template_catalog(
    survey_level: int = 3,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    schema = template_discoverer.ensure_canonical_schema(tenant_id)
    sections = []
    for sec in schema.ordered_sections():
        grp = _section_group(sec.id)
        hint = ", ".join(sec.keywords[:6]) if sec.keywords else sec.title
        sections.append({
            "code": sec.id,
            "group": grp,
            "title": sec.title,
            "hint": hint,
        })
    labels = {k: v for k, v in _GROUP_LABELS.items() if any(s["group"] == k for s in sections)}
    return {
        "survey_level": max(1, min(3, survey_level)),
        "sections": sections,
        "group_labels": labels,
    }


@router.get("/reports/{report_id}/photo-policy")
def photo_policy(report_id: str, tenant_id: str = Depends(get_current_tenant)) -> dict:
    session = _require_session(tenant_id, report_id)
    schema = template_discoverer.ensure_canonical_schema(tenant_id)
    sections = []
    for sec in schema.ordered_sections():
        sections.append({"code": sec.id, "policy": "allowed", "reason": "Photos supported in v2."})
    return {
        "report_id": session.report_id,
        "sections": sections,
        "photo_limits": {
            "max_photos_per_section": settings.max_section_photos_per_section,
            "max_photos_for_ai": settings.max_section_photos_for_ai,
        },
    }


@router.get("/reports/{report_id}/sections/{section_code}/photos")
def list_legacy_photos(
    report_id: str,
    section_code: str,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    session = _require_session(tenant_id, report_id)
    photos = _photo_rows(tenant_id, session, section_code)
    selected = sum(1 for p in photos if p.get("selected_for_ai"))
    return {
        "report_id": report_id,
        "section_code": section_code,
        "photos": photos,
        "max_photos_per_section": settings.max_section_photos_per_section,
        "max_photos_for_ai": settings.max_section_photos_for_ai,
        "selected_for_ai_count": selected,
    }


@router.post("/reports/{report_id}/sections/{section_code}/photos", status_code=201)
async def upload_legacy_photos(
    report_id: str,
    section_code: str,
    files: list[UploadFile] = File(...),
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    session = _require_session(tenant_id, report_id)
    saved = []
    for upload in files:
        ct = (upload.content_type or "image/jpeg").lower().split(";")[0].strip()
        if ct not in ALLOWED_CONTENT_TYPES:
            raise HTTPException(status_code=400, detail=f"Unsupported type: {upload.content_type}")
        data = await upload.read()
        row = photo_store.add_section_photo(
            tenant_id,
            session.draft_id,
            section_code,
            data,
            content_type=ct,
            original_filename=upload.filename or "",
        )
        saved.append({
            "photo_id": row.id,
            "original_filename": row.filename,
        })
    return {
        "report_id": report_id,
        "section_code": section_code,
        "saved": saved,
        "photos": _photo_rows(tenant_id, session, section_code),
    }


@router.put("/reports/{report_id}/sections/{section_code}/photos/ai-selection")
def legacy_photo_ai_selection(
    report_id: str,
    section_code: str,
    body: PhotoAiSelectionBody,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    session = _require_session(tenant_id, report_id)
    try:
        photo_store.set_ai_selection(tenant_id, session.draft_id, section_code, body.photo_ids)
    except ValueError as exc:
        raise HTTPException(status_code=422, detail=str(exc)) from exc
    return {
        "report_id": report_id,
        "section_code": section_code,
        "photos": _photo_rows(tenant_id, session, section_code),
    }


@router.delete("/reports/{report_id}/sections/{section_code}/photos/{photo_id}")
def delete_legacy_photo(
    report_id: str,
    section_code: str,
    photo_id: str,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    session = _require_session(tenant_id, report_id)
    if not photo_store.delete_section_photo(tenant_id, session.draft_id, section_code, photo_id):
        raise HTTPException(status_code=404, detail="Photo not found")
    return {"deleted": True, "photo_id": photo_id}


@router.get("/reports/{report_id}/sections/{section_code}/photos/{photo_id}")
def get_legacy_photo(
    report_id: str,
    section_code: str,
    photo_id: str,
    tenant_id: str = Depends(_resolve_tenant_token),
) -> Response:
    session = _require_session(tenant_id, report_id)
    rows = photo_store.list_section_photos(tenant_id, session.draft_id, section_code)
    row = next((r for r in rows if r.id == photo_id), None)
    if row is None:
        raise HTTPException(status_code=404, detail="Photo not found")
    path = photo_store.photo_file_path(
        tenant_id, session.draft_id, section_code, photo_id, content_type=row.content_type
    )
    if not path or not path.is_file():
        raise HTTPException(status_code=404, detail="Photo file missing")
    return Response(content=path.read_bytes(), media_type=row.content_type)


@router.post("/reports/{report_id}/generate")
async def generate_report_legacy(
    report_id: str,
    body: LegacyGenerateBody,
    background_tasks: BackgroundTasks,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    from backend.core.section_mapper import estimate_active_sections_from_generate_body

    session = _require_session(tenant_id, report_id)
    active = estimate_active_sections_from_generate_body(
        body,
        tenant_id=tenant_id,
        draft_id=session.draft_id,
    )
    total = len(active) if active else 1
    session.status = "generating"
    session.generation_section_total = total
    session.generation_started_at = time.time()
    session.error_message = None
    save_session(session)
    background_tasks.add_task(_run_generation, tenant_id, report_id, body)
    return {
        "report_id": report_id,
        "status": "generating",
        "message": "Generation started (v2 reference mapping).",
    }


@router.get("/reports/{report_id}/status")
def report_status(report_id: str, tenant_id: str = Depends(get_current_tenant)) -> dict:
    session = _require_session(tenant_id, report_id)
    sections_saved = len(session.sections_payload)
    sections_with_text = sum(
        1 for p in session.sections_payload.values() if (p.get("text") or "").strip()
    )
    elapsed = None
    if session.generation_started_at and session.status == "generating":
        elapsed = round(time.time() - session.generation_started_at, 1)
    return {
        "report_id": report_id,
        "status": session.status if session.status != "idle" else "complete",
        "created_at": session.created_at,
        "updated_at": session.updated_at,
        "error_message": session.error_message,
        "survey_level": session.survey_level,
        "sections_saved": sections_saved,
        "sections_with_text": sections_with_text,
        "generation_section_total": session.generation_section_total,
        "generation_elapsed_seconds": elapsed,
    }


@router.post("/reports/{report_id}/generation/cancel")
def cancel_generation(report_id: str, tenant_id: str = Depends(get_current_tenant)) -> dict:
    session = _require_session(tenant_id, report_id)
    if session.status == "generating":
        session.status = "failed"
        session.error_message = "Cancelled by user (timeout)."
        save_session(session)
    return {"report_id": report_id, "status": session.status}


@router.get("/reports/{report_id}/sections")
def get_sections(report_id: str, tenant_id: str = Depends(get_current_tenant)) -> dict:
    session = _require_session(tenant_id, report_id)
    return {"report_id": report_id, "sections": session.sections_payload}


@router.patch("/reports/{report_id}/sections/{section_code}")
def patch_section(
    report_id: str,
    section_code: str,
    body: SectionTextPatch,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    session = _require_session(tenant_id, report_id)
    payload = session.sections_payload.get(section_code, {
        "text": "",
        "confidence": 0.5,
        "provenance": [],
        "mode": "generate",
    })
    payload["text"] = body.text
    session.sections_payload[section_code] = payload
    save_session(session)
    return payload


@router.get("/reports/{report_id}/export")
def export_report(
    report_id: str,
    format: str = "docx",
    tenant_id: str = Depends(get_current_tenant),
) -> Response:
    if format.lower() != "docx":
        raise HTTPException(status_code=400, detail="Only format=docx is supported.")
    session = _require_session(tenant_id, report_id)
    schema = template_discoverer.ensure_canonical_schema(tenant_id)

    from backend.core.report_assembler import from_sections_payload

    result = from_sections_payload(
        tenant_id,
        schema,
        session.sections_payload,
        property_type=session.property_type,
        tenure=session.tenure,
    )
    from backend.core.photo_store import draft_section_photo_paths

    photo_paths = draft_section_photo_paths(
        tenant_id, session.draft_id, schema.section_ids()
    )
    data = to_docx(result, schema, section_photo_paths=photo_paths, include_footer=True)
    filename = f"RICS_Report_{report_id[:8]}.docx"
    return Response(
        content=data,
        media_type="application/vnd.openxmlformats-officedocument.wordprocessingml.document",
        headers={"Content-Disposition": f'attachment; filename="{filename}"'},
    )


# ── Optional stubs (document manager / similarity — off main path) ───────────


@router.get("/documents")
def list_documents_route(
    limit: int = 500,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    docs = list_documents(tenant_id)
    rows = sorted(docs.values(), key=lambda d: d.filename)[:limit]
    return {
        "documents": [
            {
                "document_id": d.document_id,
                "filename": d.filename,
                "status": d.status,
                "created_at": document_library.document_created_at_iso(d),
                "file_size": d.file_size or None,
                "document_purpose": "reference",
                "error": d.error,
            }
            for d in rows
        ],
        "reingest_running": document_library.is_reingest_running(tenant_id),
        "reingest_progress": document_library.reingest_progress(tenant_id),
    }


@router.delete("/documents/{document_id}")
def delete_document_route(
    document_id: str,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    session_reports = [
        s
        for s in _reports_generating_for_document(tenant_id, document_id)
    ]
    if session_reports:
        raise HTTPException(
            status_code=409,
            detail=(
                "A report is still generating from this document. "
                "Wait for it to finish before deleting the source file."
            ),
        )
    try:
        removed = document_library.remove_reference_document(tenant_id, document_id)
    except KeyError as exc:
        raise HTTPException(status_code=404, detail="Document not found") from exc
    return {
        "document_id": document_id,
        "deleted": True,
        "chunks_removed": removed,
    }


@router.post("/documents/{document_id}/reingest")
def reingest_one_document(
    document_id: str,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    if _reports_generating_for_document(tenant_id, document_id):
        raise HTTPException(
            status_code=409,
            detail="A report is still generating from this document. Wait before re-ingesting.",
        )
    try:
        result = document_library.schedule_reingest_reference_document(
            tenant_id, document_id
        )
    except KeyError as exc:
        raise HTTPException(status_code=404, detail="Document not found") from exc
    except FileNotFoundError:
        return {
            "queued": 0,
            "skipped_missing_file": 1,
            "detail": "Source file is no longer on disk; cannot re-ingest.",
        }
    return result


@router.post("/documents/reingest")
def reingest_all_route(tenant_id: str = Depends(get_current_tenant)) -> dict:
    active_doc_ids = {
        doc_id
        for doc_id in _all_generating_document_ids(tenant_id)
    }
    return document_library.schedule_reingest_all_documents(
        tenant_id,
        skip_document_ids=active_doc_ids,
    )


def _reports_dir(tenant_id: str) -> Path:
    from backend.core.report_session import _reports_dir as rd  # noqa: PLC0415

    return rd(tenant_id)


def _reports_generating_for_document(tenant_id: str, document_id: str) -> list[str]:
    out: list[str] = []
    reports_dir = _reports_dir(tenant_id)
    if not reports_dir.is_dir():
        return out
    import json

    for path in reports_dir.glob("*.json"):
        try:
            data = json.loads(path.read_text(encoding="utf-8"))
        except Exception:  # noqa: BLE001
            continue
        if data.get("status") != "generating":
            continue
        doc_ids = set(data.get("document_ids") or [])
        if data.get("primary_document_id"):
            doc_ids.add(data["primary_document_id"])
        if document_id in doc_ids:
            out.append(str(data.get("report_id") or path.stem))
    return out


def _all_generating_document_ids(tenant_id: str) -> set[str]:
    ids: set[str] = set()
    reports_dir = _reports_dir(tenant_id)
    if not reports_dir.is_dir():
        return ids
    import json

    for path in reports_dir.glob("*.json"):
        try:
            data = json.loads(path.read_text(encoding="utf-8"))
        except Exception:  # noqa: BLE001
            continue
        if data.get("status") != "generating":
            continue
        ids.update(data.get("document_ids") or [])
        if data.get("primary_document_id"):
            ids.add(data["primary_document_id"])
    return ids


@router.post("/content/similar")
def content_similar(
    body: SimilarContentBody,
    tenant_id: str = Depends(get_current_tenant),
) -> dict:
    return scan_similar_content(
        tenant_id,
        text=body.text,
        section_code=body.section_code,
        peer_sections=body.peer_sections,
        limit=body.limit,
        min_relevance_percent=body.min_relevance_percent,
        exclude_document_ids=body.exclude_document_ids,
    )


@router.post("/route-notes")
async def route_notes_route(body: RouteNotesBody) -> dict:
    """Route raw note lines into canonical L3 section buckets (server-side cascade)."""
    lines = [ln.strip() for ln in (body.lines or []) if (ln or "").strip()]
    routed, unmatched = route_lines_to_l3_sections(lines)
    return {
        "routed": routed,
        "unmatched": unmatched,
        "line_count": len(lines),
        "routed_line_count": sum(len(v) for v in routed.values()),
    }


@router.post("/extract-notes")
async def extract_notes_route(file: UploadFile = File(...)) -> dict:
    suffix = Path(file.filename or "").suffix.lower()
    if suffix not in {".docx", ".pdf", ".txt"}:
        raise HTTPException(
            status_code=422,
            detail="Unsupported file type. Upload a .docx, .pdf, or .txt notes file.",
        )
    data = await file.read(settings.max_notes_extract_bytes + 1)
    if len(data) > settings.max_notes_extract_bytes:
        raise HTTPException(status_code=413, detail="Notes file too large.")
    try:
        return extract_notes_from_upload(data, file.filename or "unknown")
    except ValueError as exc:
        raise HTTPException(status_code=422, detail=str(exc)) from exc
    except Exception as exc:  # noqa: BLE001
        raise HTTPException(status_code=500, detail=f"Failed to parse file: {exc}") from exc