hatchet-sdk 1.14.1__py3-none-any.whl → 1.14.3__py3-none-any.whl

hatchet_sdk/__init__.py

@@ -11,9 +11,9 @@ from hatchet_sdk.clients.listeners.run_event_listener import (
     StepRunEventType,
     WorkflowRunEventType,
 )
-from hatchet_sdk.clients.rest.models.accept_invite_request import AcceptInviteRequest
 
 # import models into sdk package
+from hatchet_sdk.clients.rest.models.accept_invite_request import AcceptInviteRequest
 from hatchet_sdk.clients.rest.models.api_error import APIError
 from hatchet_sdk.clients.rest.models.api_errors import APIErrors
 from hatchet_sdk.clients.rest.models.api_meta import APIMeta
@@ -166,6 +166,7 @@ from hatchet_sdk.runnables.types import (
 )
 from hatchet_sdk.runnables.workflow import TaskRunRef
 from hatchet_sdk.utils.opentelemetry import OTelAttribute
+from hatchet_sdk.utils.serde import remove_null_unicode_character
 from hatchet_sdk.worker.worker import Worker, WorkerStartOptions, WorkerStatus
 from hatchet_sdk.workflow_run import WorkflowRunRef
 
@@ -290,5 +291,6 @@ __all__ = [
     "WorkflowVersionDefinition",
     "WorkflowVersionMeta",
     "or_",
+    "remove_null_unicode_character",
     "workflow",
 ]

hatchet_sdk/clients/v1/api_client.py

@@ -1,8 +1,13 @@
+from collections.abc import Callable
 from typing import ParamSpec, TypeVar
 
+import tenacity
+
 from hatchet_sdk.clients.rest.api_client import ApiClient
 from hatchet_sdk.clients.rest.configuration import Configuration
+from hatchet_sdk.clients.rest.exceptions import ServiceException
 from hatchet_sdk.config import ClientConfig
+from hatchet_sdk.logger import logger
 from hatchet_sdk.utils.typing import JSONSerializableMapping
 
 ## Type variables to use with coroutines.
@@ -20,7 +25,7 @@ P = ParamSpec("P")
 
 
 def maybe_additional_metadata_to_kv(
-    additional_metadata: dict[str, str] | JSONSerializableMapping | None
+    additional_metadata: dict[str, str] | JSONSerializableMapping | None,
 ) -> list[str] | None:
     if not additional_metadata:
         return None
@@ -42,3 +47,24 @@ class BaseRestClient:
 
     def client(self) -> ApiClient:
         return ApiClient(self.api_config)
+
+
+def retry(func: Callable[P, R]) -> Callable[P, R]:
+    return tenacity.retry(
+        reraise=True,
+        wait=tenacity.wait_exponential_jitter(),
+        stop=tenacity.stop_after_attempt(5),
+        before_sleep=_alert_on_retry,
+        retry=tenacity.retry_if_exception(_should_retry),
+    )(func)
+
+
+def _alert_on_retry(retry_state: tenacity.RetryCallState) -> None:
+    logger.debug(
+        f"Retrying {retry_state.fn}: attempt "
+        f"{retry_state.attempt_number} ended with: {retry_state.outcome}",
+    )
+
+
+def _should_retry(ex: BaseException) -> bool:
+    return isinstance(ex, ServiceException)

hatchet_sdk/features/cron.py

@@ -19,6 +19,7 @@ from hatchet_sdk.clients.rest.models.workflow_run_order_by_direction import (
 from hatchet_sdk.clients.v1.api_client import (
     BaseRestClient,
     maybe_additional_metadata_to_kv,
+    retry,
 )
 from hatchet_sdk.utils.typing import JSONSerializableMapping
 
@@ -201,6 +202,7 @@ class CronClient(BaseRestClient):
             cron_name=cron_name,
         )
 
+    @retry
     def list(
         self,
         offset: int | None = None,
@@ -241,6 +243,7 @@ class CronClient(BaseRestClient):
                 cron_name=cron_name,
             )
 
+    @retry
     def get(self, cron_id: str) -> CronWorkflows:
         """
         Retrieve a specific workflow cron trigger by ID.

hatchet_sdk/features/filters.py

@@ -10,7 +10,7 @@ from hatchet_sdk.clients.rest.models.v1_filter_list import V1FilterList
 from hatchet_sdk.clients.rest.models.v1_update_filter_request import (
     V1UpdateFilterRequest,
 )
-from hatchet_sdk.clients.v1.api_client import BaseRestClient
+from hatchet_sdk.clients.v1.api_client import BaseRestClient, retry
 from hatchet_sdk.utils.typing import JSONSerializableMapping
 
 
@@ -41,6 +41,7 @@ class FiltersClient(BaseRestClient):
         """
         return await asyncio.to_thread(self.list, limit, offset, workflow_ids, scopes)
 
+    @retry
     def list(
         self,
         limit: int | None = None,
@@ -67,6 +68,7 @@ class FiltersClient(BaseRestClient):
                 scopes=scopes,
             )
 
+    @retry
     def get(
         self,
         filter_id: str,

hatchet_sdk/features/logs.py

@@ -3,7 +3,7 @@ import asyncio
 from hatchet_sdk.clients.rest.api.log_api import LogApi
 from hatchet_sdk.clients.rest.api_client import ApiClient
 from hatchet_sdk.clients.rest.models.v1_log_line_list import V1LogLineList
-from hatchet_sdk.clients.v1.api_client import BaseRestClient
+from hatchet_sdk.clients.v1.api_client import BaseRestClient, retry
 
 
 class LogsClient(BaseRestClient):
@@ -14,6 +14,7 @@ class LogsClient(BaseRestClient):
     def _la(self, client: ApiClient) -> LogApi:
         return LogApi(client)
 
+    @retry
     def list(self, task_run_id: str) -> V1LogLineList:
         """
         List log lines for a given task run.

hatchet_sdk/features/metrics.py

@@ -12,6 +12,7 @@ from hatchet_sdk.clients.rest.models.workflow_run_status import WorkflowRunStatu
 from hatchet_sdk.clients.v1.api_client import (
     BaseRestClient,
     maybe_additional_metadata_to_kv,
+    retry,
 )
 from hatchet_sdk.utils.typing import JSONSerializableMapping
 
@@ -27,6 +28,7 @@ class MetricsClient(BaseRestClient):
     def _ta(self, client: ApiClient) -> TenantApi:
         return TenantApi(client)
 
+    @retry
     def get_workflow_metrics(
         self,
         workflow_id: str,
@@ -66,6 +68,7 @@ class MetricsClient(BaseRestClient):
             self.get_workflow_metrics, workflow_id, status, group_key
         )
 
+    @retry
     def get_queue_metrics(
         self,
         workflow_ids: list[str] | None = None,
@@ -105,6 +108,7 @@ class MetricsClient(BaseRestClient):
             self.get_queue_metrics, workflow_ids, additional_metadata
         )
 
+    @retry
     def get_task_metrics(self) -> TenantStepRunQueueMetrics:
         """
         Retrieve queue metrics

hatchet_sdk/features/runs.py

@@ -1,10 +1,15 @@
 import asyncio
+from collections.abc import AsyncIterator
 from datetime import datetime, timedelta, timezone
 from typing import TYPE_CHECKING, Literal, overload
+from warnings import warn
 
 from pydantic import BaseModel, model_validator
 
-from hatchet_sdk.clients.listeners.run_event_listener import RunEventListenerClient
+from hatchet_sdk.clients.listeners.run_event_listener import (
+    RunEventListenerClient,
+    StepRunEventType,
+)
 from hatchet_sdk.clients.listeners.workflow_listener import PooledWorkflowRunListener
 from hatchet_sdk.clients.rest.api.task_api import TaskApi
 from hatchet_sdk.clients.rest.api.workflow_runs_api import WorkflowRunsApi
@@ -13,6 +18,7 @@ from hatchet_sdk.clients.rest.models.v1_cancel_task_request import V1CancelTaskR
 from hatchet_sdk.clients.rest.models.v1_replay_task_request import V1ReplayTaskRequest
 from hatchet_sdk.clients.rest.models.v1_task_filter import V1TaskFilter
 from hatchet_sdk.clients.rest.models.v1_task_status import V1TaskStatus
+from hatchet_sdk.clients.rest.models.v1_task_summary import V1TaskSummary
 from hatchet_sdk.clients.rest.models.v1_task_summary_list import V1TaskSummaryList
 from hatchet_sdk.clients.rest.models.v1_trigger_workflow_run_request import (
     V1TriggerWorkflowRunRequest,
@@ -21,8 +27,11 @@ from hatchet_sdk.clients.rest.models.v1_workflow_run_details import V1WorkflowRu
 from hatchet_sdk.clients.v1.api_client import (
     BaseRestClient,
     maybe_additional_metadata_to_kv,
+    retry,
 )
 from hatchet_sdk.config import ClientConfig
+from hatchet_sdk.utils.aio import gather_max_concurrency
+from hatchet_sdk.utils.datetimes import partition_date_range
 from hatchet_sdk.utils.typing import JSONSerializableMapping
 
 if TYPE_CHECKING:
@@ -110,6 +119,7 @@ class RunsClient(BaseRestClient):
     def _ta(self, client: ApiClient) -> TaskApi:
         return TaskApi(client)
 
+    @retry
     def get(self, workflow_run_id: str) -> V1WorkflowRunDetails:
         """
         Get workflow run details for a given workflow run ID.
@@ -129,6 +139,7 @@ class RunsClient(BaseRestClient):
         """
         return await asyncio.to_thread(self.get, workflow_run_id)
 
+    @retry
     def get_status(self, workflow_run_id: str) -> V1TaskStatus:
         """
         Get workflow run status for a given workflow run ID.
@@ -148,6 +159,152 @@ class RunsClient(BaseRestClient):
         """
         return await asyncio.to_thread(self.get_status, workflow_run_id)
 
+    @retry
+    def list_with_pagination(
+        self,
+        since: datetime | None = None,
+        only_tasks: bool = False,
+        offset: int | None = None,
+        limit: int | None = None,
+        statuses: list[V1TaskStatus] | None = None,
+        until: datetime | None = None,
+        additional_metadata: dict[str, str] | None = None,
+        workflow_ids: list[str] | None = None,
+        worker_id: str | None = None,
+        parent_task_external_id: str | None = None,
+        triggering_event_external_id: str | None = None,
+    ) -> list[V1TaskSummary]:
+        """
+        List task runs according to a set of filters, paginating through days
+
+        :param since: The start time for filtering task runs.
+        :param only_tasks: Whether to only list task runs.
+        :param offset: The offset for pagination.
+        :param limit: The maximum number of task runs to return.
+        :param statuses: The statuses to filter task runs by.
+        :param until: The end time for filtering task runs.
+        :param additional_metadata: Additional metadata to filter task runs by.
+        :param workflow_ids: The workflow IDs to filter task runs by.
+        :param worker_id: The worker ID to filter task runs by.
+        :param parent_task_external_id: The parent task external ID to filter task runs by.
+        :param triggering_event_external_id: The event id that triggered the task run.
+
+        :return: A list of task runs matching the specified filters.
+        """
+
+        date_ranges = partition_date_range(
+            since=since or datetime.now(tz=timezone.utc) - timedelta(days=1),
+            until=until or datetime.now(tz=timezone.utc),
+        )
+
+        with self.client() as client:
+            responses = [
+                self._wra(client).v1_workflow_run_list(
+                    tenant=self.client_config.tenant_id,
+                    since=s,
+                    until=u,
+                    only_tasks=only_tasks,
+                    offset=offset,
+                    limit=limit,
+                    statuses=statuses,
+                    additional_metadata=maybe_additional_metadata_to_kv(
+                        additional_metadata
+                    ),
+                    workflow_ids=workflow_ids,
+                    worker_id=worker_id,
+                    parent_task_external_id=parent_task_external_id,
+                    triggering_event_external_id=triggering_event_external_id,
+                )
+                for s, u in date_ranges
+            ]
+
+            ## Hack for uniqueness
+            run_id_to_run = {
+                run.metadata.id: run for record in responses for run in record.rows
+            }
+
+            return sorted(
+                run_id_to_run.values(),
+                key=lambda x: x.created_at,
+                reverse=True,
+            )
+
+    @retry
+    async def aio_list_with_pagination(
+        self,
+        since: datetime | None = None,
+        only_tasks: bool = False,
+        offset: int | None = None,
+        limit: int | None = None,
+        statuses: list[V1TaskStatus] | None = None,
+        until: datetime | None = None,
+        additional_metadata: dict[str, str] | None = None,
+        workflow_ids: list[str] | None = None,
+        worker_id: str | None = None,
+        parent_task_external_id: str | None = None,
+        triggering_event_external_id: str | None = None,
+    ) -> list[V1TaskSummary]:
+        """
+        List task runs according to a set of filters, paginating through days
+
+        :param since: The start time for filtering task runs.
+        :param only_tasks: Whether to only list task runs.
+        :param offset: The offset for pagination.
+        :param limit: The maximum number of task runs to return.
+        :param statuses: The statuses to filter task runs by.
+        :param until: The end time for filtering task runs.
+        :param additional_metadata: Additional metadata to filter task runs by.
+        :param workflow_ids: The workflow IDs to filter task runs by.
+        :param worker_id: The worker ID to filter task runs by.
+        :param parent_task_external_id: The parent task external ID to filter task runs by.
+        :param triggering_event_external_id: The event id that triggered the task run.
+
+        :return: A list of task runs matching the specified filters.
+        """
+
+        date_ranges = partition_date_range(
+            since=since or datetime.now(tz=timezone.utc) - timedelta(days=1),
+            until=until or datetime.now(tz=timezone.utc),
+        )
+
+        with self.client() as client:
+            coros = [
+                asyncio.to_thread(
+                    self._wra(client).v1_workflow_run_list,
+                    tenant=self.client_config.tenant_id,
+                    since=s,
+                    until=u,
+                    only_tasks=only_tasks,
+                    offset=offset,
+                    limit=limit,
+                    statuses=statuses,
+                    additional_metadata=maybe_additional_metadata_to_kv(
+                        additional_metadata
+                    ),
+                    workflow_ids=workflow_ids,
+                    worker_id=worker_id,
+                    parent_task_external_id=parent_task_external_id,
+                    triggering_event_external_id=triggering_event_external_id,
+                )
+                for s, u in date_ranges
+            ]
+
+            responses = await gather_max_concurrency(
+                *coros,
+                max_concurrency=3,
+            )
+
+            ## Hack for uniqueness
+            run_id_to_run = {
+                run.metadata.id: run for record in responses for run in record.rows
+            }
+
+            return sorted(
+                run_id_to_run.values(),
+                key=lambda x: x.created_at,
+                reverse=True,
+            )
+
     async def aio_list(
         self,
         since: datetime | None = None,
@@ -181,7 +338,7 @@ class RunsClient(BaseRestClient):
         """
         return await asyncio.to_thread(
             self.list,
-            since=since or datetime.now(tz=timezone.utc) - timedelta(days=1),
+            since=since,
             only_tasks=only_tasks,
             offset=offset,
             limit=limit,
@@ -194,6 +351,7 @@ class RunsClient(BaseRestClient):
             triggering_event_external_id=triggering_event_external_id,
         )
 
+    @retry
     def list(
         self,
         since: datetime | None = None,
@@ -225,10 +383,22 @@ class RunsClient(BaseRestClient):
 
         :return: A list of task runs matching the specified filters.
         """
+
+        since = since or datetime.now(tz=timezone.utc) - timedelta(days=1)
+        until = until or datetime.now(tz=timezone.utc)
+
+        if (until - since).days >= 7:
+            warn(
+                "Listing runs with a date range longer than 7 days may result in performance issues. "
+                "Consider using `list_with_pagination` or `aio_list_with_pagination` instead.",
+                RuntimeWarning,
+                stacklevel=2,
+            )
+
         with self.client() as client:
             return self._wra(client).v1_workflow_run_list(
                 tenant=self.client_config.tenant_id,
-                since=since or datetime.now(tz=timezone.utc) - timedelta(days=1),
+                since=since,
                 only_tasks=only_tasks,
                 offset=offset,
                 limit=limit,
@@ -376,6 +546,7 @@ class RunsClient(BaseRestClient):
         """
         return await asyncio.to_thread(self.bulk_cancel, opts)
 
+    @retry
     def get_result(self, run_id: str) -> JSONSerializableMapping:
         """
         Get the result of a workflow run by its external ID.
@@ -413,3 +584,13 @@ class RunsClient(BaseRestClient):
             workflow_run_listener=self.workflow_run_listener,
             runs_client=self,
         )
+
+    async def subscribe_to_stream(
+        self,
+        workflow_run_id: str,
+    ) -> AsyncIterator[str]:
+        ref = self.get_run_ref(workflow_run_id=workflow_run_id)
+
+        async for chunk in ref.stream():
+            if chunk.type == StepRunEventType.STEP_RUN_EVENT_TYPE_STREAM:
+                yield chunk.payload

hatchet_sdk/features/scheduled.py

@@ -21,6 +21,7 @@ from hatchet_sdk.clients.rest.models.workflow_run_order_by_direction import (
 from hatchet_sdk.clients.v1.api_client import (
     BaseRestClient,
     maybe_additional_metadata_to_kv,
+    retry,
 )
 from hatchet_sdk.utils.typing import JSONSerializableMapping
 
@@ -153,6 +154,7 @@ class ScheduledClient(BaseRestClient):
             statuses=statuses,
         )
 
+    @retry
     def list(
         self,
         offset: int | None = None,
@@ -193,6 +195,7 @@ class ScheduledClient(BaseRestClient):
                 statuses=statuses,
             )
 
+    @retry
     def get(self, scheduled_id: str) -> ScheduledWorkflows:
         """
         Retrieves a specific scheduled workflow by scheduled run trigger ID.

hatchet_sdk/features/tenant.py

@@ -3,7 +3,7 @@ import asyncio
 from hatchet_sdk.clients.rest.api.tenant_api import TenantApi
 from hatchet_sdk.clients.rest.api_client import ApiClient
 from hatchet_sdk.clients.rest.models.tenant import Tenant
-from hatchet_sdk.clients.v1.api_client import BaseRestClient
+from hatchet_sdk.clients.v1.api_client import BaseRestClient, retry
 
 
 class TenantClient(BaseRestClient):
@@ -14,6 +14,7 @@ class TenantClient(BaseRestClient):
     def _ta(self, client: ApiClient) -> TenantApi:
         return TenantApi(client)
 
+    @retry
     def get(self) -> Tenant:
         """
         Get the current tenant.

hatchet_sdk/features/workers.py

@@ -5,7 +5,7 @@ from hatchet_sdk.clients.rest.api_client import ApiClient
 from hatchet_sdk.clients.rest.models.update_worker_request import UpdateWorkerRequest
 from hatchet_sdk.clients.rest.models.worker import Worker
 from hatchet_sdk.clients.rest.models.worker_list import WorkerList
-from hatchet_sdk.clients.v1.api_client import BaseRestClient
+from hatchet_sdk.clients.v1.api_client import BaseRestClient, retry
 
 
 class WorkersClient(BaseRestClient):
@@ -16,6 +16,7 @@ class WorkersClient(BaseRestClient):
     def _wa(self, client: ApiClient) -> WorkerApi:
         return WorkerApi(client)
 
+    @retry
     def get(self, worker_id: str) -> Worker:
         """
         Get a worker by its ID.
@@ -35,6 +36,7 @@ class WorkersClient(BaseRestClient):
         """
         return await asyncio.to_thread(self.get, worker_id)
 
+    @retry
     def list(
         self,
     ) -> WorkerList:

hatchet_sdk/features/workflows.py

@@ -6,7 +6,7 @@ from hatchet_sdk.clients.rest.api_client import ApiClient
 from hatchet_sdk.clients.rest.models.workflow import Workflow
 from hatchet_sdk.clients.rest.models.workflow_list import WorkflowList
 from hatchet_sdk.clients.rest.models.workflow_version import WorkflowVersion
-from hatchet_sdk.clients.v1.api_client import BaseRestClient
+from hatchet_sdk.clients.v1.api_client import BaseRestClient, retry
 
 
 class WorkflowsClient(BaseRestClient):
@@ -31,6 +31,7 @@ class WorkflowsClient(BaseRestClient):
         """
         return await asyncio.to_thread(self.get, workflow_id)
 
+    @retry
     def get(self, workflow_id: str) -> Workflow:
         """
         Get a workflow by its ID.
@@ -41,6 +42,7 @@ class WorkflowsClient(BaseRestClient):
         with self.client() as client:
             return self._wa(client).workflow_get(workflow_id)
 
+    @retry
     def list(
         self,
         workflow_name: str | None = None,
@@ -81,6 +83,7 @@ class WorkflowsClient(BaseRestClient):
         """
         return await asyncio.to_thread(self.list, workflow_name, limit, offset)
 
+    @retry
     def get_version(
         self, workflow_id: str, version: str | None = None
     ) -> WorkflowVersion:

hatchet_sdk/runnables/workflow.py

@@ -1,6 +1,6 @@
 import asyncio
 from collections.abc import Callable
-from datetime import datetime, timedelta, timezone
+from datetime import datetime, timedelta
 from functools import cached_property
 from typing import TYPE_CHECKING, Any, Generic, TypeVar, cast, get_type_hints
 
@@ -310,9 +310,9 @@ class BaseWorkflow(Generic[TWorkflowInput]):
 
         :returns: A list of `V1TaskSummary` objects representing the runs of the workflow.
         """
-        response = self.client.runs.list(
+        return self.client.runs.list_with_pagination(
             workflow_ids=[self.id],
-            since=since or datetime.now(tz=timezone.utc) - timedelta(days=1),
+            since=since,
             only_tasks=only_tasks,
             offset=offset,
             limit=limit,
@@ -324,8 +324,6 @@ class BaseWorkflow(Generic[TWorkflowInput]):
             triggering_event_external_id=triggering_event_external_id,
         )
 
-        return response.rows
-
     async def aio_list_runs(
         self,
         since: datetime | None = None,
@@ -355,9 +353,9 @@ class BaseWorkflow(Generic[TWorkflowInput]):
 
         :returns: A list of `V1TaskSummary` objects representing the runs of the workflow.
         """
-        return await asyncio.to_thread(
-            self.list_runs,
-            since=since or datetime.now(tz=timezone.utc) - timedelta(days=1),
+        return await self.client.runs.aio_list_with_pagination(
+            workflow_ids=[self.id],
+            since=since,
             only_tasks=only_tasks,
             offset=offset,
             limit=limit,

hatchet_sdk/utils/aio.py

@@ -0,0 +1,21 @@
+import asyncio
+from collections.abc import Coroutine
+from typing import TypeVar
+
+T = TypeVar("T")
+
+
+async def gather_max_concurrency(
+    *tasks: Coroutine[None, None, T],
+    max_concurrency: int,
+) -> list[T]:
+    sem = asyncio.Semaphore(max_concurrency)
+
+    async def task_wrapper(task: Coroutine[None, None, T]) -> T:
+        async with sem:
+            return await task
+
+    return await asyncio.gather(
+        *(task_wrapper(task) for task in tasks),
+        return_exceptions=False,
+    )

hatchet_sdk/utils/datetimes.py

@@ -0,0 +1,35 @@
+from datetime import datetime, timedelta, timezone
+from typing import TypeVar
+
+T = TypeVar("T")
+R = TypeVar("R")
+
+
+def _to_utc(dt: datetime) -> datetime:
+    if not dt.tzinfo:
+        return dt.replace(tzinfo=timezone.utc)
+
+    return dt.astimezone(timezone.utc)
+
+
+def partition_date_range(
+    since: datetime, until: datetime
+) -> list[tuple[datetime, datetime]]:
+    since = _to_utc(since)
+    until = _to_utc(until)
+
+    ranges = []
+    current_start = since
+
+    while current_start < until:
+        next_day = (current_start + timedelta(days=1)).replace(
+            hour=0, minute=0, second=0, microsecond=0
+        )
+
+        current_end = min(next_day, until)
+
+        ranges.append((current_start, current_end))
+
+        current_start = next_day
+
+    return ranges

hatchet_sdk/utils/serde.py

@@ -0,0 +1,52 @@
+from typing import Any, TypeVar, cast, overload
+
+T = TypeVar("T")
+K = TypeVar("K")
+
+
+@overload
+def remove_null_unicode_character(data: str) -> str: ...
+
+
+@overload
+def remove_null_unicode_character(data: dict[K, T]) -> dict[K, T]: ...
+
+
+@overload
+def remove_null_unicode_character(data: list[T]) -> list[T]: ...
+
+
+@overload
+def remove_null_unicode_character(data: tuple[T, ...]) -> tuple[T, ...]: ...
+
+
+def remove_null_unicode_character(
+    data: str | dict[K, T] | list[T] | tuple[T, ...],
+) -> str | dict[K, T] | list[T] | tuple[T, ...]:
+    """
+    Recursively traverse a dictionary (a task's output) and remove the unicode escape sequence \\u0000 which will cause unexpected behavior in Hatchet.
+
+    Needed as Hatchet does not support \\u0000 in task outputs
+
+    :param data: The task output (a JSON-serializable dictionary or mapping)
+    :return: The same dictionary with all \\u0000 characters removed from strings, and nested dictionaries/lists processed recursively.
+    :raises TypeError: If the input is not a string, dictionary, list, or tuple.
+    """
+    if isinstance(data, str):
+        return data.replace("\u0000", "")
+
+    if isinstance(data, dict):
+        return {
+            key: remove_null_unicode_character(cast(Any, value))
+            for key, value in data.items()
+        }
+
+    if isinstance(data, list):
+        return [remove_null_unicode_character(cast(Any, item)) for item in data]
+
+    if isinstance(data, tuple):
+        return tuple(remove_null_unicode_character(cast(Any, item)) for item in data)
+
+    raise TypeError(
+        f"Unsupported type {type(data)}. Expected str, dict, list, or tuple."
+    )

hatchet_sdk/worker/runner/runner.py

@@ -4,9 +4,9 @@ import functools
 import json
 from collections.abc import Callable
 from concurrent.futures import ThreadPoolExecutor
-from contextlib import suppress
 from enum import Enum
 from multiprocessing import Queue
+from textwrap import dedent
 from threading import Thread, current_thread
 from typing import Any, Literal, cast, overload
 
@@ -49,6 +49,7 @@ from hatchet_sdk.runnables.contextvars import (
 )
 from hatchet_sdk.runnables.task import Task
 from hatchet_sdk.runnables.types import R, TWorkflowInput
+from hatchet_sdk.utils.serde import remove_null_unicode_character
 from hatchet_sdk.worker.action_listener_process import ActionEvent
 from hatchet_sdk.worker.runner.utils.capture_logs import (
     AsyncLogSender,
@@ -410,7 +411,7 @@ class Runner:
         )
 
     ## IMPORTANT: Keep this method's signature in sync with the wrapper in the OTel instrumentor
-    async def handle_start_step_run(self, action: Action) -> None:
+    async def handle_start_step_run(self, action: Action) -> Exception | None:
         action_name = action.action_id
 
         # Find the corresponding action function from the registry
@@ -444,8 +445,11 @@ class Runner:
 
             ## FIXME: Handle cancelled exceptions and other special exceptions
             ## that we don't want to suppress here
-            with suppress(Exception):
+            try:
                 await task
+            except Exception as e:
+                ## Used for the OTel instrumentor to capture exceptions
+                return e
 
         ## Once the step run completes, we need to remove the workflow spawn index
         ## so we don't leak memory
@@ -453,6 +457,8 @@ class Runner:
             async with spawn_index_lock:
                 workflow_spawn_indices.pop(action.key)
 
+        return None
+
     ## IMPORTANT: Keep this method's signature in sync with the wrapper in the OTel instrumentor
     async def handle_start_group_key_run(self, action: Action) -> Exception | None:
         action_name = action.action_id
@@ -550,21 +556,37 @@ class Runner:
             return ""
 
         if isinstance(output, BaseModel):
-            output = output.model_dump()
+            output = output.model_dump(mode="json")
 
         if not isinstance(output, dict):
             raise IllegalTaskOutputError(
                 f"Tasks must return either a dictionary or a Pydantic BaseModel which can be serialized to a JSON object. Got object of type {type(output)} instead."
             )
 
-        if output is not None:
-            try:
-                return json.dumps(output, default=str)
-            except Exception as e:
-                logger.error(f"Could not serialize output: {e}")
-                return str(output)
+        if output is None:
+            return ""
+
+        try:
+            serialized_output = json.dumps(output, default=str)
+        except Exception as e:
+            logger.error(f"Could not serialize output: {e}")
+            serialized_output = str(output)
+
+        if "\\u0000" in serialized_output:
+            raise IllegalTaskOutputError(
+                dedent(
+                    f"""
+                Task outputs cannot contain the unicode null character \\u0000
+
+                Please see this Discord thread: https://discord.com/channels/1088927970518909068/1384324576166678710/1386714014565928992
+                Relevant Postgres documentation: https://www.postgresql.org/docs/current/datatype-json.html
+
+                Use `hatchet_sdk.{remove_null_unicode_character.__name__}` to sanitize your output if you'd like to remove the character.
+                """
+                )
+            )
 
-        return ""
+        return serialized_output
 
     async def wait_for_tasks(self) -> None:
         running = len(self.tasks.keys())

{hatchet_sdk-1.14.1.dist-info → hatchet_sdk-1.14.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: hatchet-sdk
-Version: 1.14.1
+Version: 1.14.3
 Summary: 
 License: MIT
 Author: Alexander Belanger

{hatchet_sdk-1.14.1.dist-info → hatchet_sdk-1.14.3.dist-info}/RECORD

@@ -1,4 +1,4 @@
-hatchet_sdk/__init__.py,sha256=3iVKyZ4OnGxZWU0YtXQG9imlljPNBdPARqqCm6Ft03o,10640
+hatchet_sdk/__init__.py,sha256=LdT8xJ7-95gwksRZuCzuKDM5bfLx8uW8gXGPL7VMqwc,10743
 hatchet_sdk/client.py,sha256=OXb2hOJ5p7pY5QMlM4cydb4aGyf6bDdbyWQjPMVCe64,2413
 hatchet_sdk/clients/admin.py,sha256=t8GXhjMiFp9iL0ofISNxSOKlfwaOTX2iQbZfz1G0pHU,16936
 hatchet_sdk/clients/dispatcher/action_listener.py,sha256=tFknXH9iSP0OFYYVcKeDZVrcDNIz00ZQVTxSbHpbKhI,13863
@@ -229,7 +229,7 @@ hatchet_sdk/clients/rest/models/workflow_version_meta.py,sha256=TW4R7bAuYAg_LraN
 hatchet_sdk/clients/rest/models/workflow_workers_count.py,sha256=qhzqfvjjIDyARkiiLGluMIqEmqO-diHTsjlu0Doi0yg,2875
 hatchet_sdk/clients/rest/rest.py,sha256=zZHTzgl-NBdcK6XhG23m_s9RKRONGPPItzGe407s7GA,9262
 hatchet_sdk/clients/rest/tenacity_utils.py,sha256=SQuy4x9scgBpqygRrGdjP7dJz4u4VVAEbjnieid4KcU,1029
-hatchet_sdk/clients/v1/api_client.py,sha256=mJQUZ3cOxlFJiwWKK5F8jBxcpNZ7A2292HucrBqurbg,1205
+hatchet_sdk/clients/v1/api_client.py,sha256=Gs7OX47YsZ4mpJlItEg-MSeHk6wWb3iCwPQu1cffGZk,1974
 hatchet_sdk/conditions.py,sha256=CnhpkXgVXM3wc0kAX8KZQA6tp8NFAbdzAN2xFbw7Hb0,4522
 hatchet_sdk/config.py,sha256=DKOSCyOhFMx9d3Rvu5z9aImbOgZgwdNSg3XVzyVHn3g,5185
 hatchet_sdk/connection.py,sha256=oRxLs_lBRgHfE4jGLZJimr25ymlEJnK1ZXlh-CasjPo,2696
@@ -255,16 +255,16 @@ hatchet_sdk/contracts/workflows_pb2.py,sha256=daEsUwZnlDQ5GGLJ8WHgLdI1Tgr3lBXxGV
 hatchet_sdk/contracts/workflows_pb2.pyi,sha256=WJ3b45pWvoNmmWTWjBJt61IiAoVn61F62AG5OrRsnd8,15538
 hatchet_sdk/contracts/workflows_pb2_grpc.py,sha256=2V8E72DlJx5qlH2yiQpVCu5cQbKUba5X7T1yNrQDF_s,10819
 hatchet_sdk/exceptions.py,sha256=DG-mS0wZiB-4Pnyr-BgY-LRrAEAdgP2nqQlIheU99t4,2646
-hatchet_sdk/features/cron.py,sha256=42QJiSOXI9xAJw_d3nKhg0f_LLnwGQaLxBZ5JPhoT_w,9690
-hatchet_sdk/features/filters.py,sha256=a75EDS-gLguPEs31eVsn48mvczS2Robqju4Wr3o5B2E,6414
-hatchet_sdk/features/logs.py,sha256=OcmgtmNyqFJI03_5ncuSy6M-Ho7AVTa8hnO0CDE3wi4,1172
-hatchet_sdk/features/metrics.py,sha256=TzAEB4Ogmgcq-EB7lEWQ9V8y-15d23ZuhAgPH6It92Y,4519
+hatchet_sdk/features/cron.py,sha256=k6Y-JJBPaf2Dtx-fwvNA2j7lTzHLBZpwVMA_u-p6Lvw,9723
+hatchet_sdk/features/filters.py,sha256=n6PPeRiqd5SOFlcx8V2strUaCGma9JPRAOLx44XpC0o,6443
+hatchet_sdk/features/logs.py,sha256=H_vQnOqiN5q_wQWVoOGAJp7eOPKFYZsLJ1Hb63LriRA,1190
+hatchet_sdk/features/metrics.py,sha256=mUz51fQsrmyp_UUoZ5KxSrOmUl2mEGoYYHRuXGxl_7w,4563
 hatchet_sdk/features/rate_limits.py,sha256=eh55Z3w75cYUthqTyoWmNxj_6tN3rjebMKm3of-vxv0,2155
-hatchet_sdk/features/runs.py,sha256=mmarF1kXFVWz6ayoYBAzZTkXBQVNb2Xq60d_atnkj-U,16120
-hatchet_sdk/features/scheduled.py,sha256=bp5h6QP8B9keOsCTG1laB2GWIaXQ7ylBSNdsRFD3VU8,8906
-hatchet_sdk/features/tenant.py,sha256=vU6buEKVPCydpgrHFsQ_gbKgO5lRmlZG2ypsT7-O4S8,868
-hatchet_sdk/features/workers.py,sha256=vD6j7GCttu0fm23_XmBMdE0IuX4mUbL0adgMoC8Sk_E,2571
-hatchet_sdk/features/workflows.py,sha256=15MSYNIjlN1Ilk8sHq_DjLS7XuqlvcAMFuAdFLdjPGY,4012
+hatchet_sdk/features/runs.py,sha256=8geS5l1CU3GCLudFORmCvXyjVVcgNKsNAPV9mLcI8sU,23004
+hatchet_sdk/features/scheduled.py,sha256=t7YA9CoJrzBhH82ChTSFWaTF_dyoC9i1O_wf9ywsphc,8939
+hatchet_sdk/features/tenant.py,sha256=xkhh5mRKCWbunk_S1iBmGR-DYR-F4mjxk8jLyYUyzNE,886
+hatchet_sdk/features/workers.py,sha256=DVdno28RmtlfhMJUkaPcOMHNKXCPV0RFrXtLqV6zWyE,2600
+hatchet_sdk/features/workflows.py,sha256=a_0UwX_Y-6DX8TcQ7yuiugWHvKzQmMqafKanq4HEiSg,4052
 hatchet_sdk/hatchet.py,sha256=rOuE7YPqu4setE_zGbmgvb2WN-q8HNXTDFYcp9N0Wj8,25759
 hatchet_sdk/labels.py,sha256=nATgxWE3lFxRTnfISEpoIRLGbMfAZsHF4lZTuG4Mfic,182
 hatchet_sdk/logger.py,sha256=5uOr52T4mImSQm1QvWT8HvZFK5WfPNh3Y1cBQZRFgUQ,333
@@ -276,11 +276,14 @@ hatchet_sdk/runnables/action.py,sha256=zrVHpyzIQ9XZgWwY69b_6uhZd53An4trRoLd9b3os
 hatchet_sdk/runnables/contextvars.py,sha256=T2LWiXhcSyQYJY_-pfqMjDNjf6PdtDwyXyCZ6zIyWK0,929
 hatchet_sdk/runnables/task.py,sha256=VEYabAl38U9L_igSDgZSzUL7-c-sX6YFRP0wsqGTadU,6714
 hatchet_sdk/runnables/types.py,sha256=QFayOJ_Oi8tOYI6Sjl9bwjftM96QZh9XIlfFnSNgEXI,4359
-hatchet_sdk/runnables/workflow.py,sha256=AVYMg_QTN9YkddFuOBa8iq9oQinRPdHo9UJsgI2TWHQ,49327
+hatchet_sdk/runnables/workflow.py,sha256=SxxNhv1SAMh4Rw68_fyQJh3Su-e_zscPmILo3ghzzaU,49225
 hatchet_sdk/token.py,sha256=KjIiInwG5Kqd_FO4BSW1x_5Uc7PFbnzIVJqr50-ZldE,779
+hatchet_sdk/utils/aio.py,sha256=cu1rD_UZkShtfzi7iXMYwBBaCRdxJQTdUC0_mf8nU2E,499
 hatchet_sdk/utils/backoff.py,sha256=6B5Rb5nLKw_TqqgpJMYjIBV1PTTtbOMRZCveisVhg_I,353
+hatchet_sdk/utils/datetimes.py,sha256=vIZNEX8tt-bknaIuTmoLEmKVt18dBjClH3urYtCJAys,775
 hatchet_sdk/utils/opentelemetry.py,sha256=64TVwCLrUzEmcL2BUNPV_QubfiR5jajOZtVeGYLnEEA,1226
 hatchet_sdk/utils/proto_enums.py,sha256=v2gp_ZmIhPxURVXwz5lscllXwZXDl5XGXeL6gezw3o0,1241
+hatchet_sdk/utils/serde.py,sha256=d2iypdK2UQCPA19NgYa0Tr7CTbk25KPbaCMqcqN3CYk,1645
 hatchet_sdk/utils/timedelta_to_expression.py,sha256=YujnBnGn7lxtkUdKIeqmOiN_ZCGBpRPjCCSzcD3jxzA,644
 hatchet_sdk/utils/typing.py,sha256=FgYnZyJSoRjNVFodxlI9gn0X8ST1KFed7xfUynIxa2U,978
 hatchet_sdk/v0/__init__.py,sha256=r3Q7l2RsLgdIkK2jjiz7-JJpD1T_Zy--Oa9MN5n_yEs,9654
@@ -507,11 +510,11 @@ hatchet_sdk/v0/workflow_run.py,sha256=jsEZprXshrSV7i_TtL5uoCL03D18zQ3NeJCq7mp97D
 hatchet_sdk/worker/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 hatchet_sdk/worker/action_listener_process.py,sha256=Xzzn1dDFJrqnC9HBsh3fYI8lfpOD4Ecze47qmm_XUWE,12923
 hatchet_sdk/worker/runner/run_loop_manager.py,sha256=GeILClNXaDbsjXCQb0bBdgeyAwZGem2JdaH0t6wz__I,4053
-hatchet_sdk/worker/runner/runner.py,sha256=5dOLFJIEWHed2sQxSeWmqMFWVViYTnfvk70n1FRLoU4,20770
+hatchet_sdk/worker/runner/runner.py,sha256=Btz7wd0DQ-5aPEZLN3T5X5__8DjuxN6H4yQEVX0eZiI,21678
 hatchet_sdk/worker/runner/utils/capture_logs.py,sha256=DKw6boqVsSCM1XvBWYuc833MZxCdSpMxg3l4aAqKPyw,3465
 hatchet_sdk/worker/worker.py,sha256=nDuRo_LishRuOCTnDonc7G7qeOoW93nRHGd-fQOE_bs,16541
 hatchet_sdk/workflow_run.py,sha256=KcylcqRwKADtnzOTjoiVr1vdr7qTZFtDeD5aRS6A4Y8,2823
-hatchet_sdk-1.14.1.dist-info/METADATA,sha256=Eygcnag5XEsRLK-kC1LDbF5L3YJmLIazhyLFTOZeLBY,3636
-hatchet_sdk-1.14.1.dist-info/WHEEL,sha256=FMvqSimYX_P7y0a7UY-_Mc83r5zkBZsCYPm7Lr0Bsq4,88
-hatchet_sdk-1.14.1.dist-info/entry_points.txt,sha256=Un_76pcLse-ZGBlwebhQpnTPyQrripeHW8J7qmEpGOk,1400
-hatchet_sdk-1.14.1.dist-info/RECORD,,
+hatchet_sdk-1.14.3.dist-info/METADATA,sha256=3D-GMkNjhRG37vaE8W4lXoG1MMu_2MpnKcAwK5elTkY,3636
+hatchet_sdk-1.14.3.dist-info/WHEEL,sha256=FMvqSimYX_P7y0a7UY-_Mc83r5zkBZsCYPm7Lr0Bsq4,88
+hatchet_sdk-1.14.3.dist-info/entry_points.txt,sha256=Un_76pcLse-ZGBlwebhQpnTPyQrripeHW8J7qmEpGOk,1400
+hatchet_sdk-1.14.3.dist-info/RECORD,,