fix: Update push_data and user_data annotation with JsonSerializable instead of Any

pyproject.toml

@@ -45,7 +45,7 @@ dependencies = [
     "pydantic>=2.11.0",
     "pyee>=9.0.0",
     "tldextract>=5.1.0",
-    "typing-extensions>=4.1.0",
+    "typing-extensions>=4.10.0",
     "yarl>=1.18.0",
 ]
 

src/crawlee/_types.py

@@ -27,9 +27,7 @@
     from crawlee.storage_clients import StorageClient
     from crawlee.storages import KeyValueStore
 
-    # Workaround for https://github.com/pydantic/pydantic/issues/9445
-    J = TypeVar('J', bound='JsonSerializable')
-    JsonSerializable = list[J] | dict[str, J] | str | bool | int | float | None
+    JsonSerializable = dict[str, 'JsonSerializable'] | list['JsonSerializable'] | str | int | float | bool | None
 else:
     from pydantic import JsonValue as JsonSerializable
 
@@ -198,7 +196,7 @@ class PushDataKwargs(TypedDict):
 
 
 class PushDataFunctionCall(PushDataKwargs):
-    data: list[dict[str, Any]] | dict[str, Any]
+    data: Sequence[Mapping[str, JsonSerializable]] | Mapping[str, JsonSerializable]
     dataset_id: str | None
     dataset_name: str | None
     dataset_alias: str | None
@@ -300,7 +298,7 @@ async def add_requests(
 
     async def push_data(
         self,
-        data: list[dict[str, Any]] | dict[str, Any],
+        data: Sequence[Mapping[str, JsonSerializable]] | Mapping[str, JsonSerializable],
         dataset_id: str | None = None,
         dataset_name: str | None = None,
         dataset_alias: str | None = None,
@@ -392,7 +390,7 @@ def __call__(
         selector: str | None = None,
         attribute: str | None = None,
         label: str | None = None,
-        user_data: dict[str, Any] | None = None,
+        user_data: dict[str, JsonSerializable] | None = None,
         transform_request_function: Callable[[RequestOptions], RequestOptions | RequestTransformAction] | None = None,
         rq_id: str | None = None,
         rq_name: str | None = None,
@@ -417,7 +415,7 @@ def __call__(
         selector: str | None = None,
         attribute: str | None = None,
         label: str | None = None,
-        user_data: dict[str, Any] | None = None,
+        user_data: dict[str, JsonSerializable] | None = None,
         transform_request_function: Callable[[RequestOptions], RequestOptions | RequestTransformAction] | None = None,
         requests: Sequence[str | Request] | None = None,
         rq_id: str | None = None,
@@ -465,7 +463,7 @@ def __call__(
         selector: str = 'a',
         attribute: str = 'href',
         label: str | None = None,
-        user_data: dict[str, Any] | None = None,
+        user_data: dict[str, JsonSerializable] | None = None,
         transform_request_function: Callable[[RequestOptions], RequestOptions | RequestTransformAction] | None = None,
         **kwargs: Unpack[EnqueueLinksKwargs],
     ) -> Coroutine[None, None, list[Request]]:
@@ -543,7 +541,7 @@ class PushDataFunction(Protocol):
 
     def __call__(
         self,
-        data: list[dict[str, Any]] | dict[str, Any],
+        data: Sequence[Mapping[str, JsonSerializable]] | Mapping[str, JsonSerializable],
         dataset_id: str | None = None,
         dataset_name: str | None = None,
         dataset_alias: str | None = None,

src/crawlee/_utils/file.py

@@ -10,12 +10,12 @@
 from typing import TYPE_CHECKING, overload
 
 if TYPE_CHECKING:
-    from collections.abc import AsyncIterator
+    from collections.abc import AsyncIterator, Mapping
     from typing import Any, TextIO
 
     from typing_extensions import Unpack
 
-    from crawlee._types import ExportDataCsvKwargs, ExportDataJsonKwargs
+    from crawlee._types import ExportDataCsvKwargs, ExportDataJsonKwargs, JsonSerializable
 
 if sys.platform == 'win32':
 
@@ -150,7 +150,7 @@ async def atomic_write(
 
 
 async def export_json_to_stream(
-    iterator: AsyncIterator[dict[str, Any]],
+    iterator: AsyncIterator[Mapping[str, JsonSerializable]],
     dst: TextIO,
     **kwargs: Unpack[ExportDataJsonKwargs],
 ) -> None:
@@ -159,7 +159,7 @@ async def export_json_to_stream(
 
 
 async def export_csv_to_stream(
-    iterator: AsyncIterator[dict[str, Any]],
+    iterator: AsyncIterator[Mapping[str, JsonSerializable]],
     dst: TextIO,
     **kwargs: Unpack[ExportDataCsvKwargs],
 ) -> None:

src/crawlee/crawlers/_abstract_http/_abstract_http_crawler.py

@@ -4,7 +4,7 @@
 import logging
 from abc import ABC
 from datetime import timedelta
-from typing import TYPE_CHECKING, Any, Generic
+from typing import TYPE_CHECKING, Generic
 
 from more_itertools import partition
 from pydantic import ValidationError
@@ -26,7 +26,7 @@
     from typing_extensions import Unpack
 
     from crawlee import RequestTransformAction
-    from crawlee._types import BasicCrawlingContext, EnqueueLinksKwargs, ExtractLinksFunction
+    from crawlee._types import BasicCrawlingContext, EnqueueLinksKwargs, ExtractLinksFunction, JsonSerializable
 
     from ._abstract_http_parser import AbstractHttpParser
 
@@ -200,7 +200,7 @@ async def extract_links(
             selector: str = 'a',
             attribute: str = 'href',
             label: str | None = None,
-            user_data: dict[str, Any] | None = None,
+            user_data: dict[str, JsonSerializable] | None = None,
             transform_request_function: Callable[[RequestOptions], RequestOptions | RequestTransformAction]
             | None = None,
             **kwargs: Unpack[EnqueueLinksKwargs],

src/crawlee/crawlers/_basic/_basic_crawler.py

@@ -80,7 +80,7 @@
 
 if TYPE_CHECKING:
     import re
-    from collections.abc import Iterator
+    from collections.abc import Iterator, Mapping
     from contextlib import AbstractAsyncContextManager
 
     from crawlee._types import (
@@ -941,7 +941,7 @@ async def export_data(
 
     async def _push_data(
         self,
-        data: list[dict[str, Any]] | dict[str, Any],
+        data: Sequence[Mapping[str, JsonSerializable]] | Mapping[str, JsonSerializable],
         dataset_id: str | None = None,
         dataset_name: str | None = None,
         dataset_alias: str | None = None,
@@ -1015,7 +1015,7 @@ async def enqueue_links(
             selector: str | None = None,
             attribute: str | None = None,
             label: str | None = None,
-            user_data: dict[str, Any] | None = None,
+            user_data: dict[str, JsonSerializable] | None = None,
             transform_request_function: Callable[[RequestOptions], RequestOptions | RequestTransformAction]
             | None = None,
             requests: Sequence[str | Request] | None = None,

src/crawlee/crawlers/_playwright/_playwright_crawler.py

@@ -53,6 +53,7 @@
         HttpHeaders,
         HttpMethod,
         HttpPayload,
+        JsonSerializable,
     )
     from crawlee.browsers._types import BrowserType
 
@@ -384,7 +385,7 @@ async def extract_links(
             selector: str = 'a',
             attribute: str = 'href',
             label: str | None = None,
-            user_data: dict | None = None,
+            user_data: dict[str, JsonSerializable] | None = None,
             transform_request_function: Callable[[RequestOptions], RequestOptions | RequestTransformAction]
             | None = None,
             **kwargs: Unpack[EnqueueLinksKwargs],

src/crawlee/sessions/_models.py

@@ -13,6 +13,8 @@
     computed_field,
 )
 
+from crawlee._types import JsonSerializable
+
 from ._cookies import CookieParam
 from ._session import Session
 
@@ -24,7 +26,7 @@ class SessionModel(BaseModel):
 
     id: Annotated[str, Field(alias='id')]
     max_age: Annotated[timedelta, Field(alias='maxAge')]
-    user_data: Annotated[dict, Field(alias='userData')]
+    user_data: Annotated[dict[str, JsonSerializable], Field(alias='userData')]
     max_error_score: Annotated[float, Field(alias='maxErrorScore')]
     error_score_decrement: Annotated[float, Field(alias='errorScoreDecrement')]
     created_at: Annotated[datetime, Field(alias='createdAt')]

src/crawlee/sessions/_session.py

@@ -11,8 +11,10 @@
 from crawlee.sessions._cookies import CookieParam, SessionCookies
 
 if TYPE_CHECKING:
+    from collections.abc import Mapping, MutableMapping
     from http.cookiejar import CookieJar
 
+    from crawlee._types import JsonSerializable
     from crawlee.sessions._models import SessionModel
 
 logger = getLogger(__name__)
@@ -36,7 +38,7 @@ def __init__(
         *,
         id: str | None = None,
         max_age: timedelta = timedelta(minutes=50),
-        user_data: dict | None = None,
+        user_data: Mapping[str, JsonSerializable] | None = None,
         max_error_score: float = 3.0,
         error_score_decrement: float = 0.5,
         created_at: datetime | None = None,
@@ -63,7 +65,7 @@ def __init__(
         """
         self._id = id or crypto_random_object_id(length=10)
         self._max_age = max_age
-        self._user_data = user_data or {}
+        self._user_data: dict[str, JsonSerializable] = dict(user_data) if user_data is not None else {}
         self._max_error_score = max_error_score
         self._error_score_decrement = error_score_decrement
         self._created_at = created_at or datetime.now(timezone.utc)
@@ -117,7 +119,7 @@ def id(self) -> str:
         return self._id
 
     @property
-    def user_data(self) -> dict:
+    def user_data(self) -> MutableMapping[str, JsonSerializable]:
         """Get the user data."""
         return self._user_data
 

src/crawlee/storage_clients/_base/_dataset_client.py

@@ -4,12 +4,20 @@
 from typing import TYPE_CHECKING
 
 if TYPE_CHECKING:
-    from collections.abc import AsyncIterator
-    from typing import Any
+    from collections.abc import AsyncIterator, Mapping, Sequence
 
+    from typing_extensions import TypeIs
+
+    from crawlee._types import JsonSerializable
     from crawlee.storage_clients.models import DatasetItemsListPage, DatasetMetadata
 
 
+def _is_list_of_items(
+    data: Sequence[Mapping[str, JsonSerializable]] | Mapping[str, JsonSerializable],
+) -> TypeIs[Sequence[Mapping[str, JsonSerializable]]]:
+    return isinstance(data, list)
+
+
 class DatasetClient(ABC):
     """An abstract class for dataset storage clients.
 
@@ -42,7 +50,7 @@ async def purge(self) -> None:
         """
 
     @abstractmethod
-    async def push_data(self, data: list[Any] | dict[str, Any]) -> None:
+    async def push_data(self, data: Sequence[Mapping[str, JsonSerializable]] | Mapping[str, JsonSerializable]) -> None:
         """Push data to the dataset.
 
         The backend method for the `Dataset.push_data` call.
@@ -82,7 +90,7 @@ async def iterate_items(
         unwind: list[str] | None = None,
         skip_empty: bool = False,
         skip_hidden: bool = False,
-    ) -> AsyncIterator[dict[str, Any]]:
+    ) -> AsyncIterator[Mapping[str, JsonSerializable]]:
         """Iterate over the dataset items with filtering options.
 
         The backend method for the `Dataset.iterate_items` call.
@@ -91,3 +99,9 @@ async def iterate_items(
         raise NotImplementedError
         if False:
             yield {}
+
+    @staticmethod
+    def _is_list_of_items(
+        data: Sequence[Mapping[str, JsonSerializable]] | Mapping[str, JsonSerializable],
+    ) -> TypeIs[Sequence[Mapping[str, JsonSerializable]]]:
+        return isinstance(data, list)

src/crawlee/storage_clients/_file_system/_dataset_client.py

@@ -3,6 +3,7 @@
 import asyncio
 import json
 import shutil
+from collections.abc import Mapping
 from datetime import datetime, timezone
 from logging import getLogger
 from pathlib import Path
@@ -12,14 +13,15 @@
 from typing_extensions import Self, override
 
 from crawlee._consts import METADATA_FILENAME
+from crawlee._types import JsonSerializable
 from crawlee._utils.crypto import crypto_random_object_id
 from crawlee._utils.file import atomic_write, json_dumps
 from crawlee._utils.raise_if_too_many_kwargs import raise_if_too_many_kwargs
 from crawlee.storage_clients._base import DatasetClient
 from crawlee.storage_clients.models import DatasetItemsListPage, DatasetMetadata
 
 if TYPE_CHECKING:
-    from collections.abc import AsyncIterator
+    from collections.abc import AsyncIterator, Sequence
 
     from crawlee.configuration import Configuration
 
@@ -220,10 +222,10 @@ async def purge(self) -> None:
             )
 
     @override
-    async def push_data(self, data: list[dict[str, Any]] | dict[str, Any]) -> None:
+    async def push_data(self, data: Sequence[Mapping[str, JsonSerializable]] | Mapping[str, JsonSerializable]) -> None:
         async with self._lock:
             new_item_count = self._metadata.item_count
-            if isinstance(data, list):
+            if self._is_list_of_items(data):
                 for item in data:
                     new_item_count += 1
                     await self._push_item(item, new_item_count)
@@ -304,7 +306,7 @@ async def get_data(
             selected_files = selected_files[:limit]
 
         # Read and parse each data file.
-        items = list[dict[str, Any]]()
+        items = list[Mapping[str, JsonSerializable]]()
         for file_path in selected_files:
             try:
                 file_content = await asyncio.to_thread(file_path.read_text, encoding='utf-8')
@@ -441,7 +443,7 @@ async def _update_metadata(
         data = await json_dumps(self._metadata.model_dump())
         await atomic_write(self.path_to_metadata, data)
 
-    async def _push_item(self, item: dict[str, Any], item_id: int) -> None:
+    async def _push_item(self, item: Mapping[str, JsonSerializable], item_id: int) -> None:
         """Push a single item to the dataset.
 
         This method writes the item as a JSON file with a zero-padded numeric filename