app/db.py

from __future__ import annotations

import json
import sqlite3
from datetime import datetime, timezone
from pathlib import Path
from typing import Any, Iterator

from app.config import Settings
from app.schemas import MatchHit, ReceiptExtract, ReceiptStatus, to_cents

try:
    import sqlite_vec
    from sqlite_vec import serialize_float32
except ImportError:  # pragma: no cover
    sqlite_vec = None
    serialize_float32 = None  # type: ignore[assignment]


class VecLoadError(RuntimeError):
    pass


class EmbedIndexError(RuntimeError):
    pass


def _utc_now() -> str:
    return datetime.now(timezone.utc).replace(microsecond=0).isoformat()


def connect(path: Path) -> sqlite3.Connection:
    path.parent.mkdir(parents=True, exist_ok=True)
    con = sqlite3.connect(str(path))
    con.row_factory = sqlite3.Row
    con.execute("PRAGMA foreign_keys = ON")
    if sqlite_vec is None:
        raise VecLoadError("sqlite-vec is not installed")
    try:
        con.enable_load_extension(True)
        sqlite_vec.load(con)
        con.enable_load_extension(False)
    except Exception as exc:
        con.close()
        raise VecLoadError(f"sqlite-vec load failed: {exc}") from exc
    return con


def _create_vec_table(con: sqlite3.Connection, name: str, pk: str, dim: int) -> None:
    ddl = (
        f"CREATE VIRTUAL TABLE IF NOT EXISTS {name} USING vec0("
        f"{pk} INTEGER PRIMARY KEY, embedding float[{dim}] distance_metric=cosine)"
    )
    try:
        con.execute(ddl)
    except sqlite3.OperationalError:
        con.execute(
            f"CREATE VIRTUAL TABLE IF NOT EXISTS {name} USING vec0("
            f"{pk} INTEGER PRIMARY KEY, embedding float[{dim}])"
        )


def init_schema(con: sqlite3.Connection, settings: Settings) -> None:
    con.executescript(
        """
        CREATE TABLE IF NOT EXISTS meta (
            key TEXT PRIMARY KEY,
            value TEXT NOT NULL
        );
        CREATE TABLE IF NOT EXISTS receipts (
            id INTEGER PRIMARY KEY AUTOINCREMENT,
            source_path TEXT NOT NULL,
            sha256 TEXT NOT NULL UNIQUE,
            status TEXT NOT NULL,
            doc_kind TEXT,
            category TEXT,
            vendor TEXT,
            receipt_date TEXT,
            tax_cents INTEGER,
            total_cents INTEGER,
            currency TEXT,
            ocr_text TEXT,
            extract_json TEXT,
            error TEXT,
            created_at TEXT NOT NULL,
            updated_at TEXT NOT NULL
        );
        CREATE INDEX IF NOT EXISTS idx_receipts_category ON receipts(category);
        CREATE INDEX IF NOT EXISTS idx_receipts_vendor ON receipts(vendor);
        CREATE INDEX IF NOT EXISTS idx_receipts_date ON receipts(receipt_date);
        CREATE TABLE IF NOT EXISTS line_items (
            id INTEGER PRIMARY KEY AUTOINCREMENT,
            receipt_id INTEGER NOT NULL REFERENCES receipts(id) ON DELETE CASCADE,
            description TEXT NOT NULL,
            qty REAL,
            unit_price_cents INTEGER,
            amount_cents INTEGER,
            sku TEXT,
            match_catalog_id INTEGER,
            match_score REAL,
            match_status TEXT
        );
        CREATE TABLE IF NOT EXISTS catalog (
            id INTEGER PRIMARY KEY AUTOINCREMENT,
            sku TEXT,
            vendor TEXT,
            description TEXT NOT NULL,
            size TEXT,
            unit_price_cents INTEGER,
            metadata_json TEXT
        );
        """
    )
    _create_vec_table(con, "receipt_vec", "receipt_id", settings.embed_dim)
    _create_vec_table(con, "catalog_vec", "catalog_id", settings.embed_dim)
    _check_or_set_meta(con, settings)
    con.commit()


def _meta(con: sqlite3.Connection, key: str) -> str | None:
    row = con.execute("SELECT value FROM meta WHERE key = ?", (key,)).fetchone()
    return None if row is None else str(row["value"])


def _check_or_set_meta(con: sqlite3.Connection, settings: Settings) -> None:
    stored_model = _meta(con, "embed_model")
    stored_dim = _meta(con, "embed_dim")
    if stored_model is None:
        con.execute(
            "INSERT INTO meta(key, value) VALUES ('embed_model', ?), ('embed_dim', ?)",
            (settings.embed_model, str(settings.embed_dim)),
        )
        return
    if stored_model != settings.embed_model or stored_dim != str(settings.embed_dim):
        raise EmbedIndexError(
            f"index is {stored_model} dim={stored_dim}; config is "
            f"{settings.embed_model} dim={settings.embed_dim}. Never mix embedding models."
        )


def open_db(settings: Settings) -> sqlite3.Connection:
    con = connect(settings.db_path)
    init_schema(con, settings)
    return con


def get_by_sha(con: sqlite3.Connection, sha256: str) -> sqlite3.Row | None:
    return con.execute("SELECT * FROM receipts WHERE sha256 = ?", (sha256,)).fetchone()


def insert_receipt(
    con: sqlite3.Connection,
    *,
    source_path: str,
    sha256: str,
    status: ReceiptStatus,
    extract: ReceiptExtract | None = None,
    ocr_text: str | None = None,
    error: str | None = None,
) -> int:
    now = _utc_now()
    cur = con.execute(
        """
        INSERT INTO receipts (
            source_path, sha256, status, doc_kind, category, vendor, receipt_date,
            tax_cents, total_cents, currency, ocr_text, extract_json, error,
            created_at, updated_at
        ) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
        """,
        (
            source_path,
            sha256,
            status.value,
            None if extract is None else extract.doc_kind,
            None if extract is None else extract.category,
            None if extract is None else extract.vendor,
            None if extract is None else (extract.date.isoformat() if extract.date else None),
            None if extract is None else to_cents(extract.tax),
            None if extract is None else to_cents(extract.total),
            None if extract is None else extract.currency,
            ocr_text,
            None if extract is None else extract.model_dump_json(),
            error,
            now,
            now,
        ),
    )
    receipt_id = int(cur.lastrowid)
    if extract is not None:
        for item in extract.line_items:
            con.execute(
                """
                INSERT INTO line_items (
                    receipt_id, description, qty, unit_price_cents, amount_cents, sku
                ) VALUES (?, ?, ?, ?, ?, ?)
                """,
                (
                    receipt_id,
                    item.description,
                    item.qty,
                    to_cents(item.unit_price),
                    to_cents(item.amount),
                    item.sku,
                ),
            )
    con.commit()
    return receipt_id


def update_receipt_status(
    con: sqlite3.Connection,
    receipt_id: int,
    status: ReceiptStatus,
    *,
    error: str | None = None,
) -> None:
    con.execute(
        "UPDATE receipts SET status = ?, error = ?, updated_at = ? WHERE id = ?",
        (status.value, error, _utc_now(), receipt_id),
    )
    con.commit()


def update_extract(
    con: sqlite3.Connection,
    receipt_id: int,
    extract: ReceiptExtract,
    status: ReceiptStatus,
) -> None:
    con.execute("DELETE FROM line_items WHERE receipt_id = ?", (receipt_id,))
    con.execute(
        """
        UPDATE receipts SET
            status = ?, doc_kind = ?, category = ?, vendor = ?, receipt_date = ?,
            tax_cents = ?, total_cents = ?, currency = ?, extract_json = ?,
            error = NULL, updated_at = ?
        WHERE id = ?
        """,
        (
            status.value,
            extract.doc_kind,
            extract.category,
            extract.vendor,
            extract.date.isoformat() if extract.date else None,
            to_cents(extract.tax),
            to_cents(extract.total),
            extract.currency,
            extract.model_dump_json(),
            _utc_now(),
            receipt_id,
        ),
    )
    for item in extract.line_items:
        con.execute(
            """
            INSERT INTO line_items (
                receipt_id, description, qty, unit_price_cents, amount_cents, sku
            ) VALUES (?, ?, ?, ?, ?, ?)
            """,
            (
                receipt_id,
                item.description,
                item.qty,
                to_cents(item.unit_price),
                to_cents(item.amount),
                item.sku,
            ),
        )
    con.commit()


def delete_receipt(con: sqlite3.Connection, receipt_id: int, *, unlink_file: bool = True) -> bool:
    row = con.execute(
        "SELECT source_path FROM receipts WHERE id = ?", (receipt_id,)
    ).fetchone()
    if row is None:
        return False
    try:
        con.execute("DELETE FROM receipt_vec WHERE receipt_id = ?", (receipt_id,))
    except sqlite3.Error:
        pass
    con.execute("DELETE FROM receipts WHERE id = ?", (receipt_id,))
    con.commit()
    if unlink_file:
        path = Path(row["source_path"] or "")
        if path.is_file():
            try:
                path.unlink()
            except OSError:
                pass
    return True


def set_line_match(
    con: sqlite3.Connection,
    line_id: int,
    hit: MatchHit,
) -> None:
    con.execute(
        """
        UPDATE line_items SET match_catalog_id = ?, match_score = ?, match_status = ?
        WHERE id = ?
        """,
        (hit.catalog_id, hit.similarity, hit.band.value, line_id),
    )
    con.commit()


def list_receipts(con: sqlite3.Connection, *, status: str | None = None, limit: int = 50) -> list[sqlite3.Row]:
    if status:
        return list(
            con.execute(
                "SELECT * FROM receipts WHERE status = ? ORDER BY id DESC LIMIT ?",
                (status, limit),
            )
        )
    return list(con.execute("SELECT * FROM receipts ORDER BY id DESC LIMIT ?", (limit,)))


def get_receipt(con: sqlite3.Connection, receipt_id: int) -> sqlite3.Row | None:
    return con.execute("SELECT * FROM receipts WHERE id = ?", (receipt_id,)).fetchone()


def list_line_items(con: sqlite3.Connection, receipt_id: int) -> list[sqlite3.Row]:
    return list(
        con.execute("SELECT * FROM line_items WHERE receipt_id = ? ORDER BY id", (receipt_id,))
    )


def add_catalog_item(
    con: sqlite3.Connection,
    *,
    description: str,
    sku: str | None = None,
    vendor: str | None = None,
    size: str | None = None,
    unit_price_cents: int | None = None,
    metadata: dict[str, Any] | None = None,
) -> int:
    cur = con.execute(
        """
        INSERT INTO catalog (sku, vendor, description, size, unit_price_cents, metadata_json)
        VALUES (?, ?, ?, ?, ?, ?)
        """,
        (
            sku,
            vendor,
            description,
            size,
            unit_price_cents,
            None if metadata is None else json.dumps(metadata),
        ),
    )
    con.commit()
    return int(cur.lastrowid)


def list_catalog(con: sqlite3.Connection, limit: int = 200) -> list[sqlite3.Row]:
    return list(con.execute("SELECT * FROM catalog ORDER BY id DESC LIMIT ?", (limit,)))


def find_catalog_by_sku(con: sqlite3.Connection, sku: str) -> sqlite3.Row | None:
    return con.execute(
        "SELECT * FROM catalog WHERE sku = ? COLLATE NOCASE LIMIT 1", (sku,)
    ).fetchone()


def upsert_vector(con: sqlite3.Connection, table: str, pk_col: str, pk: int, vec: list[float]) -> None:
    if serialize_float32 is None:
        raise VecLoadError("sqlite-vec missing")
    blob = serialize_float32(vec)
    con.execute(f"DELETE FROM {table} WHERE {pk_col} = ?", (pk,))
    con.execute(
        f"INSERT INTO {table}({pk_col}, embedding) VALUES (?, ?)",
        (pk, blob),
    )
    con.commit()


def knn(
    con: sqlite3.Connection,
    table: str,
    pk_col: str,
    query: list[float],
    *,
    k: int = 5,
) -> list[tuple[int, float]]:
    if serialize_float32 is None:
        raise VecLoadError("sqlite-vec missing")
    blob = serialize_float32(query)
    rows = con.execute(
        f"""
        SELECT {pk_col} AS id, distance
        FROM {table}
        WHERE embedding MATCH ?
          AND k = ?
        """,
        (blob, k),
    ).fetchall()
    return [(int(row["id"]), float(row["distance"])) for row in rows]


def receipt_to_extract(row: sqlite3.Row) -> ReceiptExtract | None:
    raw = row["extract_json"]
    if not raw:
        return None
    return ReceiptExtract.model_validate_json(raw)


def iter_rows(rows: list[sqlite3.Row]) -> Iterator[dict[str, Any]]:
    for row in rows:
        yield dict(row)