sqlspec 0.47.0__cp314-cp314-win_amd64.whl

sqlspec/adapters/duckdb/litestar/store.py

@@ -0,0 +1,330 @@
+"""DuckDB sync session store for Litestar integration."""
+
+from datetime import datetime, timedelta, timezone
+from typing import TYPE_CHECKING
+
+from sqlspec.extensions.litestar.store import BaseSQLSpecStore
+from sqlspec.utils.sync_tools import async_
+
+if TYPE_CHECKING:
+    from sqlspec.adapters.duckdb.config import DuckDBConfig
+
+
+__all__ = ("DuckdbStore",)
+
+
+class DuckdbStore(BaseSQLSpecStore["DuckDBConfig"]):
+    """DuckDB session store using synchronous DuckDB driver.
+
+    Implements server-side session storage for Litestar using DuckDB
+    via the synchronous duckdb driver. Uses Litestar's sync_to_thread
+    utility to provide an async interface compatible with the Store protocol.
+
+    Provides efficient session management with:
+    - Sync operations wrapped for async compatibility
+    - INSERT OR REPLACE for UPSERT functionality
+    - Native TIMESTAMP type support
+    - Automatic expiration handling
+    - Efficient cleanup of expired sessions
+    - Columnar storage optimized for analytical queries
+
+    Note:
+        DuckDB is primarily designed for analytical (OLAP) workloads.
+        For high-concurrency OLTP session stores, consider PostgreSQL adapters.
+
+    Args:
+        config: DuckDBConfig instance.
+
+    Example:
+        from sqlspec.adapters.duckdb import DuckDBConfig
+        from sqlspec.adapters.duckdb.litestar.store import DuckdbStore
+
+        config = DuckDBConfig()
+        store = DuckdbStore(config)
+        await store.create_table()
+    """
+
+    __slots__ = ()
+
+    def __init__(self, config: "DuckDBConfig") -> None:
+        """Initialize DuckDB session store.
+
+        Args:
+            config: DuckDBConfig instance.
+
+        Notes:
+            Table name is read from config.extension_config["litestar"]["session_table"].
+        """
+        super().__init__(config)
+
+    def _get_create_table_sql(self) -> str:
+        """Get DuckDB CREATE TABLE SQL.
+
+        Returns:
+            SQL statement to create the sessions table with proper indexes.
+
+        Notes:
+            - Uses TIMESTAMP type for expires_at (DuckDB native datetime type)
+            - TIMESTAMP supports ISO 8601 format and direct comparisons
+            - Columnar storage makes this efficient for analytical queries
+            - DuckDB does not support partial indexes, so full index is created
+        """
+        return f"""
+        CREATE TABLE IF NOT EXISTS {self._table_name} (
+            session_id VARCHAR PRIMARY KEY,
+            data BLOB NOT NULL,
+            expires_at TIMESTAMP,
+            created_at TIMESTAMP DEFAULT NOW(),
+            updated_at TIMESTAMP DEFAULT NOW()
+        );
+        CREATE INDEX IF NOT EXISTS idx_{self._table_name}_expires_at
+        ON {self._table_name}(expires_at);
+        """
+
+    def _get_drop_table_sql(self) -> "list[str]":
+        """Get DuckDB DROP TABLE SQL statements.
+
+        Returns:
+            List of SQL statements to drop indexes and table.
+        """
+        return [f"DROP INDEX IF EXISTS idx_{self._table_name}_expires_at", f"DROP TABLE IF EXISTS {self._table_name}"]
+
+    def _datetime_to_timestamp(self, dt: "datetime | None") -> "str | None":
+        """Convert datetime to ISO 8601 string for DuckDB TIMESTAMP storage.
+
+        Args:
+            dt: Datetime to convert (must be UTC-aware).
+
+        Returns:
+            ISO 8601 formatted string, or None if dt is None.
+
+        Notes:
+            DuckDB's TIMESTAMP type accepts ISO 8601 format strings.
+            This enables efficient storage and comparison operations.
+        """
+        if dt is None:
+            return None
+        return dt.isoformat()
+
+    def _timestamp_to_datetime(self, ts: "str | datetime | None") -> "datetime | None":
+        """Convert TIMESTAMP string back to datetime.
+
+        Args:
+            ts: ISO 8601 timestamp string or datetime object.
+
+        Returns:
+            UTC-aware datetime, or None if ts is None.
+        """
+        if ts is None:
+            return None
+        if isinstance(ts, datetime):
+            if ts.tzinfo is None:
+                return ts.replace(tzinfo=timezone.utc)
+            return ts
+        dt = datetime.fromisoformat(ts)
+        if dt.tzinfo is None:
+            dt = dt.replace(tzinfo=timezone.utc)
+        return dt
+
+    def _create_table(self) -> None:
+        """Synchronous implementation of create_table."""
+        sql = self._get_create_table_sql()
+        with self._config.provide_session() as driver:
+            driver.execute_script(sql)
+        self._log_table_created()
+
+    async def create_table(self) -> None:
+        """Create the session table if it doesn't exist."""
+        await async_(self._create_table)()
+
+    def _get(self, key: str, renew_for: "int | timedelta | None" = None) -> "bytes | None":
+        """Synchronous implementation of get."""
+        sql = f"""
+        SELECT data, expires_at FROM {self._table_name}
+        WHERE session_id = ?
+        AND (expires_at IS NULL OR expires_at > CURRENT_TIMESTAMP)
+        """
+
+        with self._config.provide_connection() as conn:
+            cursor = conn.execute(sql, (key,))
+            row = cursor.fetchone()
+
+            if row is None:
+                return None
+
+            data, expires_at_str = row
+
+            if renew_for is not None and expires_at_str is not None:
+                new_expires_at = self._calculate_expires_at(renew_for)
+                new_expires_at_str = self._datetime_to_timestamp(new_expires_at)
+                if new_expires_at_str is not None:
+                    update_sql = f"""
+                    UPDATE {self._table_name}
+                    SET expires_at = ?, updated_at = NOW()
+                    WHERE session_id = ?
+                    """
+                    conn.execute(update_sql, (new_expires_at_str, key))
+                    conn.commit()
+
+            return bytes(data)
+
+    async def get(self, key: str, renew_for: "int | timedelta | None" = None) -> "bytes | None":
+        """Get a session value by key.
+
+        Args:
+            key: Session ID to retrieve.
+            renew_for: If given, renew the expiry time for this duration.
+
+        Returns:
+            Session data as bytes if found and not expired, None otherwise.
+        """
+        return await async_(self._get)(key, renew_for)
+
+    def _set(self, key: str, value: "str | bytes", expires_in: "int | timedelta | None" = None) -> None:
+        """Synchronous implementation of set.
+
+        Notes:
+            Stores expires_at as TIMESTAMP (ISO 8601 string) for DuckDB native support.
+            Uses INSERT ON CONFLICT instead of INSERT OR REPLACE to ensure all columns
+            are properly updated. created_at uses DEFAULT on insert, updated_at gets
+            current timestamp on both insert and update.
+        """
+        data = self._value_to_bytes(value)
+        expires_at = self._calculate_expires_at(expires_in)
+        expires_at_str = self._datetime_to_timestamp(expires_at)
+
+        sql = f"""
+        INSERT INTO {self._table_name} (session_id, data, expires_at)
+        VALUES (?, ?, ?)
+        ON CONFLICT (session_id)
+        DO UPDATE SET
+            data = EXCLUDED.data,
+            expires_at = EXCLUDED.expires_at,
+            updated_at = NOW()
+        """
+
+        with self._config.provide_connection() as conn:
+            conn.execute(sql, (key, data, expires_at_str))
+            conn.commit()
+
+    async def set(self, key: str, value: "str | bytes", expires_in: "int | timedelta | None" = None) -> None:
+        """Store a session value.
+
+        Args:
+            key: Session ID.
+            value: Session data.
+            expires_in: Time until expiration.
+        """
+        await async_(self._set)(key, value, expires_in)
+
+    def _delete(self, key: str) -> None:
+        """Synchronous implementation of delete."""
+        sql = f"DELETE FROM {self._table_name} WHERE session_id = ?"
+
+        with self._config.provide_connection() as conn:
+            conn.execute(sql, (key,))
+            conn.commit()
+
+    async def delete(self, key: str) -> None:
+        """Delete a session by key.
+
+        Args:
+            key: Session ID to delete.
+        """
+        await async_(self._delete)(key)
+
+    def _delete_all(self) -> None:
+        """Synchronous implementation of delete_all."""
+        sql = f"DELETE FROM {self._table_name}"
+
+        with self._config.provide_connection() as conn:
+            conn.execute(sql)
+            conn.commit()
+        self._log_delete_all()
+
+    async def delete_all(self) -> None:
+        """Delete all sessions from the store."""
+        await async_(self._delete_all)()
+
+    def _exists(self, key: str) -> bool:
+        """Synchronous implementation of exists."""
+        sql = f"""
+        SELECT 1 FROM {self._table_name}
+        WHERE session_id = ?
+        AND (expires_at IS NULL OR expires_at > CURRENT_TIMESTAMP)
+        """
+
+        with self._config.provide_connection() as conn:
+            cursor = conn.execute(sql, (key,))
+            result = cursor.fetchone()
+            return result is not None
+
+    async def exists(self, key: str) -> bool:
+        """Check if a session key exists and is not expired.
+
+        Args:
+            key: Session ID to check.
+
+        Returns:
+            True if the session exists and is not expired.
+        """
+        return await async_(self._exists)(key)
+
+    def _expires_in(self, key: str) -> "int | None":
+        """Synchronous implementation of expires_in."""
+        sql = f"""
+        SELECT expires_at FROM {self._table_name}
+        WHERE session_id = ?
+        """
+
+        with self._config.provide_connection() as conn:
+            cursor = conn.execute(sql, (key,))
+            row = cursor.fetchone()
+
+            if row is None or row[0] is None:
+                return None
+
+            expires_at_str = row[0]
+            expires_at = self._timestamp_to_datetime(expires_at_str)
+
+            if expires_at is None:
+                return None
+
+            now = datetime.now(timezone.utc)
+
+            if expires_at <= now:
+                return 0
+
+            delta = expires_at - now
+            return int(delta.total_seconds())
+
+    async def expires_in(self, key: str) -> "int | None":
+        """Get the time in seconds until the session expires.
+
+        Args:
+            key: Session ID to check.
+
+        Returns:
+            Seconds until expiration, or None if no expiry or key doesn't exist.
+        """
+        return await async_(self._expires_in)(key)
+
+    def _delete_expired(self) -> int:
+        """Synchronous implementation of delete_expired."""
+        sql = f"DELETE FROM {self._table_name} WHERE expires_at <= CURRENT_TIMESTAMP"
+
+        with self._config.provide_connection() as conn:
+            cursor = conn.execute(sql)
+            count = cursor.fetchone()
+            row_count = count[0] if count else 0
+            if row_count > 0:
+                self._log_delete_expired(row_count)
+            return row_count
+
+    async def delete_expired(self) -> int:
+        """Delete all expired sessions.
+
+        Returns:
+            Number of sessions deleted.
+        """
+        return await async_(self._delete_expired)()

sqlspec/adapters/duckdb/pool.py

@@ -0,0 +1,350 @@
+"""DuckDB connection pool with thread-local connections."""
+
+import logging
+import threading
+import time
+import uuid
+from contextlib import contextmanager, suppress
+from typing import TYPE_CHECKING, Any, Final, cast
+
+import duckdb
+
+from sqlspec.adapters.duckdb._typing import DuckDBConnection
+from sqlspec.utils.logging import POOL_LOGGER_NAME, get_logger, log_with_context
+
+if TYPE_CHECKING:
+    from collections.abc import Callable, Generator
+
+
+logger = get_logger(POOL_LOGGER_NAME)
+_ADAPTER_NAME = "duckdb"
+
+DEFAULT_MIN_POOL: Final[int] = 1
+DEFAULT_MAX_POOL: Final[int] = 4
+POOL_TIMEOUT: Final[float] = 30.0
+POOL_RECYCLE: Final[int] = 86400
+HEALTH_CHECK_INTERVAL: Final[float] = 30.0
+
+__all__ = ("DuckDBConnectionPool",)
+
+
+class DuckDBConnectionPool:
+    """Thread-local connection manager for DuckDB.
+
+    Uses thread-local storage to ensure each thread gets its own DuckDB connection,
+    preventing the thread-safety issues that cause segmentation faults when
+    multiple cursors share the same connection concurrently.
+
+    This design trades traditional pooling for thread safety, which is essential
+    for DuckDB since connections and cursors are not thread-safe.
+    """
+
+    __slots__ = (
+        "_connection_config",
+        "_connection_times",
+        "_created_connections",
+        "_extension_flags",
+        "_extensions",
+        "_health_check_interval",
+        "_is_memory_db",
+        "_lock",
+        "_on_connection_create",
+        "_pool_id",
+        "_recycle",
+        "_secrets",
+        "_thread_local",
+    )
+
+    def __init__(
+        self,
+        connection_config: "dict[str, Any]",
+        pool_recycle_seconds: int = POOL_RECYCLE,
+        health_check_interval: float = HEALTH_CHECK_INTERVAL,
+        extensions: "list[dict[str, Any]] | None" = None,
+        extension_flags: "dict[str, Any] | None" = None,
+        secrets: "list[dict[str, Any]] | None" = None,
+        on_connection_create: "Callable[[DuckDBConnection], DuckDBConnection | None] | None" = None,
+    ) -> None:
+        """Initialize the thread-local connection manager.
+
+        Args:
+            connection_config: DuckDB connection configuration
+            pool_recycle_seconds: Connection recycle time in seconds
+            health_check_interval: Seconds of idle time before running health check
+            extensions: List of extensions to install/load
+            extension_flags: Connection-level SET statements applied after creation
+            secrets: List of secrets to create
+            on_connection_create: Callback executed when connection is created
+        """
+        self._connection_config = connection_config
+        self._recycle = pool_recycle_seconds
+        self._health_check_interval = health_check_interval
+        self._extensions = extensions or []
+        self._extension_flags = extension_flags or {}
+        self._secrets = secrets or []
+        self._on_connection_create = on_connection_create
+        self._thread_local = threading.local()
+        self._lock = threading.RLock()
+        self._created_connections = 0
+        self._connection_times: dict[int, float] = {}
+        self._pool_id = str(uuid.uuid4())[:8]
+        # Track if this pool uses an in-memory database
+        # In-memory databases require connections to stay alive to preserve data
+        database = connection_config.get("database", "")
+        self._is_memory_db = database.startswith(":memory:") or database == ""
+
+    @property
+    def _database_name(self) -> str:
+        """Get sanitized database name for logging."""
+        db = self._connection_config.get("database", "")
+        if db.startswith(":memory:") or db == "":
+            return ":memory:"
+        return str(db)
+
+    def _create_connection(self) -> DuckDBConnection:
+        """Create a new DuckDB connection with extensions and secrets."""
+        connect_parameters = {}
+        config_dict = {}
+
+        for key, value in self._connection_config.items():
+            if key in {"database", "read_only"}:
+                connect_parameters[key] = value
+            else:
+                config_dict[key] = value
+
+        if config_dict:
+            connect_parameters["config"] = config_dict
+
+        connection = duckdb.connect(**connect_parameters)
+
+        self._apply_extension_flags(connection)
+
+        for ext_config in self._extensions:
+            ext_name = ext_config.get("name")
+            if not ext_name:
+                continue
+
+            install_kwargs = {}
+            if "version" in ext_config:
+                install_kwargs["version"] = ext_config["version"]
+            if "repository" in ext_config:
+                install_kwargs["repository"] = ext_config["repository"]
+            if ext_config.get("force_install", False):
+                install_kwargs["force_install"] = True
+
+            try:
+                if install_kwargs:
+                    connection.install_extension(ext_name, **install_kwargs)
+                connection.load_extension(ext_name)
+            except Exception as e:
+                log_with_context(
+                    logger,
+                    logging.DEBUG,
+                    "pool.extension.load.failed",
+                    adapter=_ADAPTER_NAME,
+                    pool_id=self._pool_id,
+                    database=self._database_name,
+                    extension=ext_name,
+                    error=str(e),
+                )
+
+        for secret_config in self._secrets:
+            secret_type = secret_config.get("secret_type")
+            secret_name = secret_config.get("name")
+            secret_value = secret_config.get("value")
+
+            if not (secret_type and secret_name and secret_value):
+                continue
+
+            value_pairs = []
+            for key, value in secret_value.items():
+                escaped_value = str(value).replace("'", "''")
+                value_pairs.append(f"'{key}' = '{escaped_value}'")
+            value_string = ", ".join(value_pairs)
+            scope_clause = ""
+            if "scope" in secret_config:
+                scope_clause = f" SCOPE '{secret_config['scope']}'"
+
+            sql = f"""
+                CREATE SECRET {secret_name} (
+                    TYPE {secret_type},
+                    {value_string}
+                ){scope_clause}
+            """
+            with suppress(Exception):
+                connection.execute(sql)
+
+        if self._on_connection_create:
+            with suppress(Exception):
+                self._on_connection_create(connection)
+
+        conn_id = id(connection)
+        with self._lock:
+            self._created_connections += 1
+            self._connection_times[conn_id] = time.time()
+
+        return connection
+
+    def _apply_extension_flags(self, connection: DuckDBConnection) -> None:
+        """Apply connection-level extension flags via SET statements."""
+
+        if not self._extension_flags:
+            return
+
+        for key, value in self._extension_flags.items():
+            if not key or not key.replace("_", "").isalnum():
+                continue
+
+            normalized = self._normalize_flag_value(value)
+            try:
+                connection.execute(f"SET {key} = {normalized}")
+            except Exception as exc:  # pragma: no cover - best-effort guard
+                log_with_context(
+                    logger,
+                    logging.DEBUG,
+                    "pool.flag.set.failed",
+                    adapter=_ADAPTER_NAME,
+                    pool_id=self._pool_id,
+                    database=self._database_name,
+                    flag=key,
+                    error=str(exc),
+                )
+
+    @staticmethod
+    def _normalize_flag_value(value: Any) -> str:
+        """Convert Python value to DuckDB SET literal."""
+
+        if isinstance(value, bool):
+            return "TRUE" if value else "FALSE"
+        if isinstance(value, (int, float)):
+            return str(value)
+        escaped = str(value).replace("'", "''")
+        return f"'{escaped}'"
+
+    def _get_thread_connection(self) -> DuckDBConnection:
+        """Get or create a connection for the current thread.
+
+        Each thread gets its own dedicated DuckDB connection to prevent
+        thread-safety issues with concurrent cursor operations.
+        """
+        thread_state = self._thread_local.__dict__
+        if "connection" not in thread_state:
+            self._thread_local.connection = self._create_connection()
+            self._thread_local.created_at = time.time()
+            self._thread_local.last_used = time.time()
+            return cast("DuckDBConnection", self._thread_local.connection)
+
+        if self._recycle > 0 and time.time() - self._thread_local.created_at > self._recycle:
+            with suppress(Exception):
+                self._thread_local.connection.close()
+            self._thread_local.connection = self._create_connection()
+            self._thread_local.created_at = time.time()
+            self._thread_local.last_used = time.time()
+            return cast("DuckDBConnection", self._thread_local.connection)
+
+        idle_time = time.time() - thread_state.get("last_used", 0)
+        if idle_time > self._health_check_interval and not self._is_connection_alive(self._thread_local.connection):
+            log_with_context(
+                logger,
+                logging.DEBUG,
+                "pool.connection.recycle",
+                adapter=_ADAPTER_NAME,
+                pool_id=self._pool_id,
+                database=self._database_name,
+                idle_seconds=round(idle_time, 1),
+                reason="failed_health_check",
+            )
+            with suppress(Exception):
+                self._thread_local.connection.close()
+            self._thread_local.connection = self._create_connection()
+            self._thread_local.created_at = time.time()
+
+        self._thread_local.last_used = time.time()
+        return cast("DuckDBConnection", self._thread_local.connection)
+
+    def _close_thread_connection(self) -> None:
+        """Close the connection for the current thread."""
+        thread_state = self._thread_local.__dict__
+        if "connection" in thread_state:
+            with suppress(Exception):
+                self._thread_local.connection.close()
+            del self._thread_local.connection
+            if "created_at" in thread_state:
+                del self._thread_local.created_at
+            if "last_used" in thread_state:
+                del self._thread_local.last_used
+
+    def _is_connection_alive(self, connection: DuckDBConnection) -> bool:
+        """Check if a connection is still alive and usable.
+
+        Args:
+            connection: Connection to check
+
+        Returns:
+            True if connection is alive, False otherwise
+        """
+        try:
+            cursor = connection.cursor()
+            cursor.close()
+        except Exception:
+            return False
+        return True
+
+    @contextmanager
+    def get_connection(self) -> "Generator[DuckDBConnection, None, None]":
+        """Get a thread-local connection.
+
+        Each thread gets its own dedicated DuckDB connection to prevent
+        thread-safety issues with concurrent cursor operations.
+
+        For file-based databases, the connection is closed when the context
+        manager exits to release DuckDB's file lock, allowing subsequent
+        connections with different configurations.
+
+        For in-memory databases, connections are kept alive to preserve data,
+        as in-memory data is lost when the last connection closes.
+
+        Yields:
+            DuckDBConnection: A thread-local connection.
+        """
+        connection = self._get_thread_connection()
+        try:
+            yield connection
+        except Exception:
+            self._close_thread_connection()
+            raise
+        else:
+            # Only close connection for file-based databases to release file locks
+            # In-memory databases need connections to stay alive to preserve data
+            if not self._is_memory_db:
+                self._close_thread_connection()
+
+    def close(self) -> None:
+        """Close the thread-local connection if it exists."""
+        self._close_thread_connection()
+
+    def size(self) -> int:
+        """Get current pool size (always 1 for thread-local)."""
+        return 1 if "connection" in self._thread_local.__dict__ else 0
+
+    def checked_out(self) -> int:
+        """Get number of checked out connections (always 0 for thread-local)."""
+        return 0
+
+    def acquire(self) -> DuckDBConnection:
+        """Acquire a thread-local connection.
+
+        Each thread gets its own dedicated DuckDB connection to prevent
+        thread-safety issues with concurrent cursor operations.
+
+        Returns:
+            DuckDBConnection: A thread-local connection
+        """
+        return self._get_thread_connection()
+
+    def release(self, connection: DuckDBConnection) -> None:
+        """Release a connection (no-op for thread-local connections).
+
+        Args:
+            connection: The connection to release (ignored)
+        """