crawlee 0.6.13b17__py3-none-any.whl → 1.1.2b7__py3-none-any.whl

crawlee/storage_clients/_sql/_request_queue_client.py

@@ -0,0 +1,720 @@
+from __future__ import annotations
+
+from collections import deque
+from datetime import datetime, timedelta, timezone
+from functools import lru_cache
+from hashlib import sha256
+from logging import getLogger
+from typing import TYPE_CHECKING, Any, cast
+
+from sqlalchemy import CursorResult, func, or_, select, update
+from sqlalchemy.exc import SQLAlchemyError
+from sqlalchemy.orm import load_only
+from typing_extensions import NotRequired, Self, override
+
+from crawlee import Request
+from crawlee._utils.crypto import crypto_random_object_id
+from crawlee.storage_clients._base import RequestQueueClient
+from crawlee.storage_clients.models import (
+    AddRequestsResponse,
+    ProcessedRequest,
+    RequestQueueMetadata,
+    UnprocessedRequest,
+)
+
+from ._client_mixin import MetadataUpdateParams, SqlClientMixin
+from ._db_models import RequestDb, RequestQueueMetadataDb, RequestQueueStateDb
+
+if TYPE_CHECKING:
+    from collections.abc import Sequence
+
+    from sqlalchemy.ext.asyncio import AsyncSession
+
+    from ._storage_client import SqlStorageClient
+
+
+logger = getLogger(__name__)
+
+
+class _QueueMetadataUpdateParams(MetadataUpdateParams):
+    """Parameters for updating queue metadata."""
+
+    new_handled_request_count: NotRequired[int]
+    new_pending_request_count: NotRequired[int]
+    new_total_request_count: NotRequired[int]
+    delta_handled_request_count: NotRequired[int]
+    delta_pending_request_count: NotRequired[int]
+    recalculate: NotRequired[bool]
+    update_had_multiple_clients: NotRequired[bool]
+
+
+class SqlRequestQueueClient(RequestQueueClient, SqlClientMixin):
+    """SQL implementation of the request queue client.
+
+    This client persists requests to a SQL database with transaction handling and
+    concurrent access safety. Requests are stored with sequence-based ordering and
+    efficient querying capabilities.
+
+    The implementation uses negative sequence numbers for forefront (high-priority) requests
+    and positive sequence numbers for regular requests, allowing for efficient single-query
+    ordering. A cache mechanism reduces database queries.
+
+    The request queue data is stored in SQL database tables following the pattern:
+    - `request_queues` table: Contains queue metadata (id, name, timestamps, request counts, multi-client flag)
+    - `request_queue_records` table: Contains individual requests with JSON data, unique keys for deduplication,
+    sequence numbers for ordering, and processing status flags
+    - `request_queue_state` table: Maintains counters for sequence numbers to ensure proper ordering of requests.
+
+    Requests are serialized to JSON for storage and maintain proper ordering through sequence
+    numbers. The implementation provides concurrent access safety through transaction
+    handling, locking mechanisms, and optimized database indexes for efficient querying.
+    """
+
+    _DEFAULT_NAME = 'default'
+    """Default dataset name used when no name is provided."""
+
+    _MAX_BATCH_FETCH_SIZE = 10
+    """Maximum number of requests to fetch from the database in a single batch operation.
+
+    Used to limit the number of requests loaded and locked for processing at once (improves efficiency and reduces
+    database load).
+    """
+
+    _METADATA_TABLE = RequestQueueMetadataDb
+    """SQLAlchemy model for request queue metadata."""
+
+    _ITEM_TABLE = RequestDb
+    """SQLAlchemy model for request items."""
+
+    _CLIENT_TYPE = 'Request queue'
+    """Human-readable client type for error messages."""
+
+    _BLOCK_REQUEST_TIME = 300
+    """Number of seconds for which a request is considered blocked in the database after being fetched for processing.
+    """
+
+    def __init__(
+        self,
+        *,
+        id: str,
+        storage_client: SqlStorageClient,
+    ) -> None:
+        """Initialize a new instance.
+
+        Preferably use the `SqlRequestQueueClient.open` class method to create a new instance.
+        """
+        super().__init__(id=id, storage_client=storage_client)
+
+        self._pending_fetch_cache: deque[Request] = deque()
+        """Cache for requests: ordered by sequence number."""
+
+        self.client_key = crypto_random_object_id(length=32)[:32]
+        """Unique identifier for this client instance."""
+
+    @classmethod
+    async def open(
+        cls,
+        *,
+        id: str | None,
+        name: str | None,
+        alias: str | None,
+        storage_client: SqlStorageClient,
+    ) -> Self:
+        """Open an existing request queue or create a new one.
+
+        This method first tries to find an existing queue by ID or name.
+        If found, it returns a client for that queue. If not found, it creates
+        a new queue with the specified parameters.
+
+        Args:
+            id: The ID of the request queue to open. Takes precedence over name.
+            name: The name of the request queue for named (global scope) storages.
+            alias: The alias of the request queue for unnamed (run scope) storages.
+            storage_client: The SQL storage client used to access the database.
+
+        Returns:
+            An instance for the opened or created request queue.
+
+        Raises:
+            ValueError: If a queue with the specified ID is not found.
+        """
+        return await cls._safely_open(
+            id=id,
+            name=name,
+            alias=alias,
+            storage_client=storage_client,
+            metadata_model=RequestQueueMetadata,
+            extra_metadata_fields={
+                'had_multiple_clients': False,
+                'handled_request_count': 0,
+                'pending_request_count': 0,
+                'total_request_count': 0,
+            },
+        )
+
+    @override
+    async def get_metadata(self) -> RequestQueueMetadata:
+        # The database is a single place of truth
+        return await self._get_metadata(RequestQueueMetadata)
+
+    @override
+    async def drop(self) -> None:
+        """Delete this request queue and all its records from the database.
+
+        This operation is irreversible. Uses CASCADE deletion to remove all related records.
+        """
+        await self._drop()
+
+        self._pending_fetch_cache.clear()
+
+    @override
+    async def purge(self) -> None:
+        """Remove all items from this dataset while keeping the dataset structure.
+
+        Resets pending_request_count and handled_request_count to 0 and deletes all records from request_queue_records
+        table.
+        """
+        await self._purge(
+            metadata_kwargs=_QueueMetadataUpdateParams(
+                update_accessed_at=True,
+                update_modified_at=True,
+                new_pending_request_count=0,
+                force=True,
+            )
+        )
+
+        # Clear recoverable state
+        self._pending_fetch_cache.clear()
+
+    @override
+    async def add_batch_of_requests(
+        self,
+        requests: Sequence[Request],
+        *,
+        forefront: bool = False,
+    ) -> AddRequestsResponse:
+        if not requests:
+            return AddRequestsResponse(processed_requests=[], unprocessed_requests=[])
+
+        # Clear empty cache since we're adding requests
+        processed_requests = []
+        unprocessed_requests = []
+        transaction_processed_requests = []
+        transaction_processed_requests_unique_keys = set()
+
+        metadata_recalculate = False
+
+        # Deduplicate requests by unique_key upfront
+        unique_requests = {}
+        unique_key_by_request_id = {}
+        for req in requests:
+            if req.unique_key not in unique_requests:
+                request_id = self._get_int_id_from_unique_key(req.unique_key)
+                unique_requests[request_id] = req
+                unique_key_by_request_id[request_id] = req.unique_key
+
+        # Get existing requests by unique keys
+        stmt = (
+            select(self._ITEM_TABLE)
+            .where(
+                self._ITEM_TABLE.request_queue_id == self._id,
+                self._ITEM_TABLE.request_id.in_(set(unique_requests.keys())),
+            )
+            .options(
+                load_only(
+                    self._ITEM_TABLE.request_id,
+                    self._ITEM_TABLE.is_handled,
+                    self._ITEM_TABLE.time_blocked_until,
+                )
+            )
+        )
+
+        async with self.get_session() as session:
+            result = await session.execute(stmt)
+            result = cast('CursorResult', result) if not isinstance(result, CursorResult) else result
+            existing_requests = {req.request_id: req for req in result.scalars()}
+            state = await self._get_state(session)
+            insert_values: list[dict] = []
+
+            for request_id, request in sorted(unique_requests.items()):
+                existing_req_db = existing_requests.get(request_id)
+                # New Request, add it
+                if existing_req_db is None:
+                    value = {
+                        'request_id': request_id,
+                        'request_queue_id': self._id,
+                        'data': request.model_dump_json(),
+                        'is_handled': False,
+                    }
+                    if forefront:
+                        value['sequence_number'] = state.forefront_sequence_counter
+                        state.forefront_sequence_counter -= 1
+                    else:
+                        value['sequence_number'] = state.sequence_counter
+                        state.sequence_counter += 1
+
+                    insert_values.append(value)
+                    metadata_recalculate = True
+                    transaction_processed_requests.append(
+                        ProcessedRequest(
+                            unique_key=request.unique_key,
+                            was_already_present=False,
+                            was_already_handled=False,
+                        )
+                    )
+                    transaction_processed_requests_unique_keys.add(request.unique_key)
+                # Already handled request, skip adding
+                elif existing_req_db and existing_req_db.is_handled:
+                    processed_requests.append(
+                        ProcessedRequest(
+                            unique_key=request.unique_key,
+                            was_already_present=True,
+                            was_already_handled=True,
+                        )
+                    )
+                # Already in progress in one of the clients
+                elif existing_req_db and existing_req_db.time_blocked_until:
+                    processed_requests.append(
+                        ProcessedRequest(
+                            unique_key=request.unique_key,
+                            was_already_present=True,
+                            was_already_handled=False,
+                        )
+                    )
+                # Request in database but not yet handled and not in progress
+                elif existing_req_db and not existing_req_db.is_handled and not existing_req_db.time_blocked_until:
+                    # Forefront request, update its sequence number
+                    if forefront:
+                        insert_values.append(
+                            {
+                                'request_queue_id': self._id,
+                                'request_id': request_id,
+                                'sequence_number': state.forefront_sequence_counter,
+                                'data': request.model_dump_json(),
+                                'is_handled': False,
+                            }
+                        )
+                        state.forefront_sequence_counter -= 1
+                        transaction_processed_requests.append(
+                            ProcessedRequest(
+                                unique_key=request.unique_key,
+                                was_already_present=True,
+                                was_already_handled=False,
+                            )
+                        )
+                        transaction_processed_requests_unique_keys.add(request.unique_key)
+                    # Regular request, keep its position
+                    else:
+                        processed_requests.append(
+                            ProcessedRequest(
+                                unique_key=request.unique_key,
+                                was_already_present=True,
+                                was_already_handled=False,
+                            )
+                        )
+                # Unexpected condition
+                else:
+                    unprocessed_requests.append(
+                        UnprocessedRequest(
+                            unique_key=request.unique_key,
+                            url=request.url,
+                            method=request.method,
+                        )
+                    )
+
+            if insert_values:
+                if forefront:
+                    # If the request already exists in the database, we update the sequence_number by shifting request
+                    # to the left.
+                    upsert_stmt = self._build_upsert_stmt(
+                        self._ITEM_TABLE,
+                        insert_values,
+                        update_columns=['sequence_number'],
+                        conflict_cols=['request_id', 'request_queue_id'],
+                    )
+                    await session.execute(upsert_stmt)
+                else:
+                    # If the request already exists in the database, we ignore this request when inserting.
+                    insert_stmt_with_ignore = self._build_insert_stmt_with_ignore(self._ITEM_TABLE, insert_values)
+                    await session.execute(insert_stmt_with_ignore)
+
+            await self._update_metadata(
+                session,
+                **_QueueMetadataUpdateParams(
+                    recalculate=metadata_recalculate,
+                    update_modified_at=True,
+                    update_accessed_at=True,
+                    force=metadata_recalculate,
+                ),
+            )
+
+            try:
+                await session.commit()
+                processed_requests.extend(transaction_processed_requests)
+            except SQLAlchemyError as e:
+                await session.rollback()
+                logger.warning(f'Failed to commit session: {e}')
+                await self._update_metadata(
+                    session, recalculate=True, update_modified_at=True, update_accessed_at=True, force=True
+                )
+                await session.commit()
+                transaction_processed_requests.clear()
+                unprocessed_requests.extend(
+                    [
+                        UnprocessedRequest(
+                            unique_key=request.unique_key,
+                            url=request.url,
+                            method=request.method,
+                        )
+                        for request in requests
+                        if request.unique_key in transaction_processed_requests_unique_keys
+                    ]
+                )
+
+        return AddRequestsResponse(
+            processed_requests=processed_requests,
+            unprocessed_requests=unprocessed_requests,
+        )
+
+    @override
+    async def get_request(self, unique_key: str) -> Request | None:
+        request_id = self._get_int_id_from_unique_key(unique_key)
+
+        stmt = select(self._ITEM_TABLE).where(
+            self._ITEM_TABLE.request_queue_id == self._id, self._ITEM_TABLE.request_id == request_id
+        )
+        async with self.get_session() as session:
+            result = await session.execute(stmt)
+            request_db = result.scalar_one_or_none()
+
+            if request_db is None:
+                logger.warning(f'Request with ID "{unique_key}" not found in the queue.')
+                return None
+
+            updated = await self._update_metadata(session, update_accessed_at=True)
+
+            # Commit updates to the metadata
+            if updated:
+                await session.commit()
+
+        return Request.model_validate_json(request_db.data)
+
+    @override
+    async def fetch_next_request(self) -> Request | None:
+        if self._pending_fetch_cache:
+            return self._pending_fetch_cache.popleft()
+
+        now = datetime.now(timezone.utc)
+        block_until = now + timedelta(seconds=self._BLOCK_REQUEST_TIME)
+        dialect = self._storage_client.get_dialect_name()
+
+        # Get available requests not blocked by another client
+        stmt = (
+            select(self._ITEM_TABLE)
+            .where(
+                self._ITEM_TABLE.request_queue_id == self._id,
+                self._ITEM_TABLE.is_handled.is_(False),
+                or_(self._ITEM_TABLE.time_blocked_until.is_(None), self._ITEM_TABLE.time_blocked_until < now),
+            )
+            .order_by(self._ITEM_TABLE.sequence_number.asc())
+            .limit(self._MAX_BATCH_FETCH_SIZE)
+        )
+
+        async with self.get_session() as session:
+            # We use the `skip_locked` database mechanism to prevent the 'interception' of requests by another client
+            if dialect == 'postgresql':
+                stmt = stmt.with_for_update(skip_locked=True)
+                result = await session.execute(stmt)
+                requests_db = result.scalars().all()
+
+                if not requests_db:
+                    return None
+
+                # All requests received have already been reserved for update with the help of `skip_locked`.
+                request_ids = {r.request_id for r in requests_db}
+
+                update_stmt = (
+                    update(self._ITEM_TABLE)
+                    .where(self._ITEM_TABLE.request_id.in_(request_ids))
+                    .values(time_blocked_until=block_until, client_key=self.client_key)
+                )
+                await session.execute(update_stmt)
+
+                blocked_ids = request_ids
+            else:
+                # For other databases, we first select the requests, then try to update them to be blocked.
+                result = await session.execute(stmt)
+                requests_db = result.scalars().all()
+
+                if not requests_db:
+                    return None
+
+                request_ids = {r.request_id for r in requests_db}
+
+                update_stmt = (
+                    update(self._ITEM_TABLE)
+                    .where(
+                        self._ITEM_TABLE.request_queue_id == self._id,
+                        self._ITEM_TABLE.request_id.in_(request_ids),
+                        self._ITEM_TABLE.is_handled.is_(False),
+                        or_(self._ITEM_TABLE.time_blocked_until.is_(None), self._ITEM_TABLE.time_blocked_until < now),
+                    )
+                    .values(time_blocked_until=block_until, client_key=self.client_key)
+                    .returning(self._ITEM_TABLE.request_id)
+                )
+
+                update_result = await session.execute(update_stmt)
+                blocked_ids = {row[0] for row in update_result.fetchall()}
+
+                if not blocked_ids:
+                    await session.rollback()
+                    return None
+
+            await self._update_metadata(session, **_QueueMetadataUpdateParams(update_accessed_at=True))
+
+            await session.commit()
+
+        requests = [Request.model_validate_json(r.data) for r in requests_db if r.request_id in blocked_ids]
+
+        if not requests:
+            return None
+
+        self._pending_fetch_cache.extend(requests[1:])
+
+        return requests[0]
+
+    @override
+    async def mark_request_as_handled(self, request: Request) -> ProcessedRequest | None:
+        request_id = self._get_int_id_from_unique_key(request.unique_key)
+
+        # Update the request's handled_at timestamp.
+        if request.handled_at is None:
+            request.handled_at = datetime.now(timezone.utc)
+
+        # Update request in Db
+        stmt = (
+            update(self._ITEM_TABLE)
+            .where(self._ITEM_TABLE.request_queue_id == self._id, self._ITEM_TABLE.request_id == request_id)
+            .values(is_handled=True, time_blocked_until=None, client_key=None, data=request.model_dump_json())
+        )
+        async with self.get_session() as session:
+            result = await session.execute(stmt)
+            result = cast('CursorResult', result) if not isinstance(result, CursorResult) else result
+
+            if result.rowcount == 0:
+                logger.warning(f'Request {request.unique_key} not found in database.')
+                return None
+
+            await self._update_metadata(
+                session,
+                **_QueueMetadataUpdateParams(
+                    delta_handled_request_count=1,
+                    delta_pending_request_count=-1,
+                    update_modified_at=True,
+                    update_accessed_at=True,
+                    force=True,
+                ),
+            )
+            await session.commit()
+        return ProcessedRequest(
+            unique_key=request.unique_key,
+            was_already_present=True,
+            was_already_handled=True,
+        )
+
+    @override
+    async def reclaim_request(
+        self,
+        request: Request,
+        *,
+        forefront: bool = False,
+    ) -> ProcessedRequest | None:
+        request_id = self._get_int_id_from_unique_key(request.unique_key)
+
+        stmt = update(self._ITEM_TABLE).where(
+            self._ITEM_TABLE.request_queue_id == self._id, self._ITEM_TABLE.request_id == request_id
+        )
+
+        async with self.get_session(with_simple_commit=True) as session:
+            state = await self._get_state(session)
+
+            # Update sequence number if changing priority
+            if forefront:
+                new_sequence = state.forefront_sequence_counter
+                state.forefront_sequence_counter -= 1
+                now = datetime.now(timezone.utc)
+                block_until = now + timedelta(seconds=self._BLOCK_REQUEST_TIME)
+                # Extend blocking for forefront request, it is considered blocked by the current client.
+                stmt = stmt.values(
+                    sequence_number=new_sequence,
+                    time_blocked_until=block_until,
+                    client_key=self.client_key,
+                    data=request.model_dump_json(),
+                )
+            else:
+                new_sequence = state.sequence_counter
+                state.sequence_counter += 1
+                stmt = stmt.values(
+                    sequence_number=new_sequence,
+                    time_blocked_until=None,
+                    client_key=None,
+                    data=request.model_dump_json(),
+                )
+
+            result = await session.execute(stmt)
+            result = cast('CursorResult', result) if not isinstance(result, CursorResult) else result
+
+            if result.rowcount == 0:
+                logger.warning(f'Request {request.unique_key} not found in database.')
+                return None
+            await self._update_metadata(
+                session, **_QueueMetadataUpdateParams(update_modified_at=True, update_accessed_at=True)
+            )
+
+        # put the forefront request at the beginning of the cache
+        if forefront:
+            self._pending_fetch_cache.appendleft(request)
+
+        return ProcessedRequest(
+            unique_key=request.unique_key,
+            was_already_present=True,
+            was_already_handled=False,
+        )
+
+    @override
+    async def is_empty(self) -> bool:
+        # Check in-memory cache for requests
+        if self._pending_fetch_cache:
+            return False
+
+        # Check database for unhandled requests
+        async with self.get_session() as session:
+            metadata_orm = await session.get(self._METADATA_TABLE, self._id)
+            if not metadata_orm:
+                raise ValueError(f'Request queue with ID "{self._id}" not found.')
+
+            empty = metadata_orm.pending_request_count == 0
+
+            updated = await self._update_metadata(
+                session,
+                **_QueueMetadataUpdateParams(
+                    update_accessed_at=True,
+                    # With multi-client access, counters may become out of sync.
+                    # If the queue is not empty, we perform a recalculation to synchronize the counters in the metadata.
+                    recalculate=not empty,
+                    update_modified_at=not empty,
+                ),
+            )
+
+            # Commit updates to the metadata
+            if updated:
+                await session.commit()
+
+        return empty
+
+    async def _get_state(self, session: AsyncSession) -> RequestQueueStateDb:
+        """Get the current state of the request queue."""
+        orm_state: RequestQueueStateDb | None = await session.get(RequestQueueStateDb, self._id)
+        if not orm_state:
+            insert_values = {'request_queue_id': self._id}
+            # Create a new state if it doesn't exist
+            # This is a safeguard against race conditions where multiple clients might try to create the state
+            # simultaneously.
+            insert_stmt = self._build_insert_stmt_with_ignore(RequestQueueStateDb, insert_values)
+            await session.execute(insert_stmt)
+            await session.flush()
+            orm_state = await session.get(RequestQueueStateDb, self._id)
+            if not orm_state:
+                raise RuntimeError(f'Failed to create or retrieve state for queue {self._id}')
+        return orm_state
+
+    def _specific_update_metadata(
+        self,
+        new_handled_request_count: int | None = None,
+        new_pending_request_count: int | None = None,
+        new_total_request_count: int | None = None,
+        delta_handled_request_count: int | None = None,
+        delta_pending_request_count: int | None = None,
+        *,
+        recalculate: bool = False,
+        update_had_multiple_clients: bool = False,
+        **_kwargs: dict[str, Any],
+    ) -> dict[str, Any]:
+        """Update the request queue metadata in the database.
+
+        Args:
+            session: The SQLAlchemy session to use for database operations.
+            new_handled_request_count: If provided, update the handled_request_count to this value.
+            new_pending_request_count: If provided, update the pending_request_count to this value.
+            new_total_request_count: If provided, update the total_request_count to this value.
+            delta_handled_request_count: If provided, add this value to the handled_request_count.
+            delta_pending_request_count: If provided, add this value to the pending_request_count.
+            recalculate: If True, recalculate the pending_request_count, and total_request_count on request table.
+            update_had_multiple_clients: If True, set had_multiple_clients to True.
+        """
+        values_to_set: dict[str, Any] = {}
+
+        if update_had_multiple_clients:
+            values_to_set['had_multiple_clients'] = True
+
+        if new_handled_request_count is not None:
+            values_to_set['handled_request_count'] = new_handled_request_count
+        elif delta_handled_request_count is not None:
+            values_to_set['handled_request_count'] = (
+                self._METADATA_TABLE.handled_request_count + delta_handled_request_count
+            )
+
+        if new_pending_request_count is not None:
+            values_to_set['pending_request_count'] = new_pending_request_count
+        elif delta_pending_request_count is not None:
+            values_to_set['pending_request_count'] = (
+                self._METADATA_TABLE.pending_request_count + delta_pending_request_count
+            )
+
+        if new_total_request_count is not None:
+            values_to_set['total_request_count'] = new_total_request_count
+
+        if recalculate:
+            stmt = (
+                update(self._METADATA_TABLE)
+                .where(self._METADATA_TABLE.request_queue_id == self._id)
+                .values(
+                    pending_request_count=(
+                        select(func.count())
+                        .select_from(self._ITEM_TABLE)
+                        .where(self._ITEM_TABLE.request_queue_id == self._id, self._ITEM_TABLE.is_handled.is_(False))
+                        .scalar_subquery()
+                    ),
+                    total_request_count=(
+                        select(func.count())
+                        .select_from(self._ITEM_TABLE)
+                        .where(self._ITEM_TABLE.request_queue_id == self._id)
+                        .scalar_subquery()
+                    ),
+                    handled_request_count=(
+                        select(func.count())
+                        .select_from(self._ITEM_TABLE)
+                        .where(self._ITEM_TABLE.request_queue_id == self._id, self._ITEM_TABLE.is_handled.is_(True))
+                        .scalar_subquery()
+                    ),
+                )
+            )
+
+            values_to_set['custom_stmt'] = stmt
+
+        return values_to_set
+
+    @staticmethod
+    @lru_cache(maxsize=10000)
+    def _get_int_id_from_unique_key(unique_key: str) -> int:
+        """Generate a deterministic integer ID for a unique_key.
+
+        Args:
+            unique_key: Unique key to be used to generate ID.
+
+        Returns:
+            An integer ID based on the unique_key.
+        """
+        hashed_key = sha256(unique_key.encode('utf-8')).hexdigest()
+        name_length = 15
+        return int(hashed_key[:name_length], 16)