feat(api): update API spec from langfuse/langfuse c974e07

langfuse/api/comments/client.py

@@ -1,5 +1,6 @@
 # This file was auto-generated by Fern from our API Definition.
 
+import datetime as dt
 import typing
 
 from ..commons.types.comment import Comment
@@ -36,6 +37,7 @@ def create(
         object_id: str,
         content: str,
         author_user_id: typing.Optional[str] = OMIT,
+        object_start_time: typing.Optional[dt.datetime] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> CreateCommentResponse:
         """
@@ -58,6 +60,9 @@ def create(
         author_user_id : typing.Optional[str]
             The id of the user who created the comment. Must be a member of the organization that owns the project, otherwise an error will be thrown.
 
+        object_start_time : typing.Optional[dt.datetime]
+            The start time of the referenced object (for observations, the observation's start time). Optional performance hint - when provided, Langfuse narrows the lookup to validate the reference faster. It only affects speed - an incorrect or omitted value never changes the result.
+
         request_options : typing.Optional[RequestOptions]
             Request-specific configuration.
 
@@ -90,6 +95,7 @@ def create(
             object_id=object_id,
             content=content,
             author_user_id=author_user_id,
+            object_start_time=object_start_time,
             request_options=request_options,
         )
         return _response.data
@@ -221,6 +227,7 @@ async def create(
         object_id: str,
         content: str,
         author_user_id: typing.Optional[str] = OMIT,
+        object_start_time: typing.Optional[dt.datetime] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> CreateCommentResponse:
         """
@@ -243,6 +250,9 @@ async def create(
         author_user_id : typing.Optional[str]
             The id of the user who created the comment. Must be a member of the organization that owns the project, otherwise an error will be thrown.
 
+        object_start_time : typing.Optional[dt.datetime]
+            The start time of the referenced object (for observations, the observation's start time). Optional performance hint - when provided, Langfuse narrows the lookup to validate the reference faster. It only affects speed - an incorrect or omitted value never changes the result.
+
         request_options : typing.Optional[RequestOptions]
             Request-specific configuration.
 
@@ -283,6 +293,7 @@ async def main() -> None:
             object_id=object_id,
             content=content,
             author_user_id=author_user_id,
+            object_start_time=object_start_time,
             request_options=request_options,
         )
         return _response.data

langfuse/api/comments/raw_client.py

@@ -1,5 +1,6 @@
 # This file was auto-generated by Fern from our API Definition.
 
+import datetime as dt
 import typing
 from json.decoder import JSONDecodeError
 
@@ -34,6 +35,7 @@ def create(
         object_id: str,
         content: str,
         author_user_id: typing.Optional[str] = OMIT,
+        object_start_time: typing.Optional[dt.datetime] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> HttpResponse[CreateCommentResponse]:
         """
@@ -56,6 +58,9 @@ def create(
         author_user_id : typing.Optional[str]
             The id of the user who created the comment. Must be a member of the organization that owns the project, otherwise an error will be thrown.
 
+        object_start_time : typing.Optional[dt.datetime]
+            The start time of the referenced object (for observations, the observation's start time). Optional performance hint - when provided, Langfuse narrows the lookup to validate the reference faster. It only affects speed - an incorrect or omitted value never changes the result.
+
         request_options : typing.Optional[RequestOptions]
             Request-specific configuration.
 
@@ -72,6 +77,7 @@ def create(
                 "objectId": object_id,
                 "content": content,
                 "authorUserId": author_user_id,
+                "objectStartTime": object_start_time,
             },
             request_options=request_options,
             omit=OMIT,
@@ -398,6 +404,7 @@ async def create(
         object_id: str,
         content: str,
         author_user_id: typing.Optional[str] = OMIT,
+        object_start_time: typing.Optional[dt.datetime] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> AsyncHttpResponse[CreateCommentResponse]:
         """
@@ -420,6 +427,9 @@ async def create(
         author_user_id : typing.Optional[str]
             The id of the user who created the comment. Must be a member of the organization that owns the project, otherwise an error will be thrown.
 
+        object_start_time : typing.Optional[dt.datetime]
+            The start time of the referenced object (for observations, the observation's start time). Optional performance hint - when provided, Langfuse narrows the lookup to validate the reference faster. It only affects speed - an incorrect or omitted value never changes the result.
+
         request_options : typing.Optional[RequestOptions]
             Request-specific configuration.
 
@@ -436,6 +446,7 @@ async def create(
                 "objectId": object_id,
                 "content": content,
                 "authorUserId": author_user_id,
+                "objectStartTime": object_start_time,
             },
             request_options=request_options,
             omit=OMIT,

langfuse/api/comments/types/create_comment_request.py

@@ -1,5 +1,6 @@
 # This file was auto-generated by Fern from our API Definition.
 
+import datetime as dt
 import typing
 
 import pydantic
@@ -42,6 +43,13 @@ class CreateCommentRequest(UniversalBaseModel):
     The id of the user who created the comment. Must be a member of the organization that owns the project, otherwise an error will be thrown.
     """
 
+    object_start_time: typing_extensions.Annotated[
+        typing.Optional[dt.datetime], FieldMetadata(alias="objectStartTime")
+    ] = pydantic.Field(default=None)
+    """
+    The start time of the referenced object (for observations, the observation's start time). Optional performance hint - when provided, Langfuse narrows the lookup to validate the reference faster. It only affects speed - an incorrect or omitted value never changes the result.
+    """
+
     model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(
         extra="allow", frozen=True
     )

langfuse/api/ingestion/client.py

@@ -29,7 +29,7 @@ def with_raw_response(self) -> RawIngestionClient:
         return self._raw_client
 
     @typing_extensions.deprecated(
-        "On Langfuse Cloud, Langfuse v3 is deprecated and this endpoint will be removed on November 16, 2026. Send data via the OpenTelemetry endpoint at `POST /api/public/otel/v1/traces` instead, see the [OpenTelemetry integration docs](https://langfuse.com/integrations/native/opentelemetry). Self-hosted deployments are unaffected by this date; the endpoint becomes unavailable when they upgrade to Langfuse v4.",
+        "On Langfuse Cloud, Langfuse v3 is deprecated and v4-only write mode begins on November 16, 2026. This endpoint is never shut down; it continues to accept score events. Trace and observation events fail only in v4-only write mode, not in dual or legacy mode. Always prefer upgrading to the current Python and JS SDKs. If you use custom auto-instrumentation, only then send data via the OpenTelemetry endpoint at `POST /api/public/otel/v1/traces` (for example with curl); see the [OpenTelemetry integration docs](https://langfuse.com/integrations/native/opentelemetry). Self-hosted deployments are unaffected by this date; they reject trace and observation events only in v4-only write mode, not dual or legacy.",
         category=None,
     )
     def batch(
@@ -42,7 +42,7 @@ def batch(
         """
         **Legacy endpoint for batch ingestion for Langfuse Observability.**
 
-        -> Please use the OpenTelemetry endpoint (`/api/public/otel/v1/traces`). Learn more: https://langfuse.com/integrations/native/opentelemetry
+        This endpoint is never shut down. Trace and observation events are rejected only in v4-only write mode (not dual or legacy); score events continue to be accepted. Always prefer upgrading to the current Python and JS SDKs. If you use custom auto-instrumentation, only then send traces via the OpenTelemetry endpoint (`POST /api/public/otel/v1/traces`), for example with curl. Learn more: https://langfuse.com/integrations/native/opentelemetry
 
         Within each batch, there can be multiple events.
         Each event has a type, an id, a timestamp, metadata and a body.
@@ -134,7 +134,7 @@ def with_raw_response(self) -> AsyncRawIngestionClient:
         return self._raw_client
 
     @typing_extensions.deprecated(
-        "On Langfuse Cloud, Langfuse v3 is deprecated and this endpoint will be removed on November 16, 2026. Send data via the OpenTelemetry endpoint at `POST /api/public/otel/v1/traces` instead, see the [OpenTelemetry integration docs](https://langfuse.com/integrations/native/opentelemetry). Self-hosted deployments are unaffected by this date; the endpoint becomes unavailable when they upgrade to Langfuse v4.",
+        "On Langfuse Cloud, Langfuse v3 is deprecated and v4-only write mode begins on November 16, 2026. This endpoint is never shut down; it continues to accept score events. Trace and observation events fail only in v4-only write mode, not in dual or legacy mode. Always prefer upgrading to the current Python and JS SDKs. If you use custom auto-instrumentation, only then send data via the OpenTelemetry endpoint at `POST /api/public/otel/v1/traces` (for example with curl); see the [OpenTelemetry integration docs](https://langfuse.com/integrations/native/opentelemetry). Self-hosted deployments are unaffected by this date; they reject trace and observation events only in v4-only write mode, not dual or legacy.",
         category=None,
     )
     async def batch(
@@ -147,7 +147,7 @@ async def batch(
         """
         **Legacy endpoint for batch ingestion for Langfuse Observability.**
 
-        -> Please use the OpenTelemetry endpoint (`/api/public/otel/v1/traces`). Learn more: https://langfuse.com/integrations/native/opentelemetry
+        This endpoint is never shut down. Trace and observation events are rejected only in v4-only write mode (not dual or legacy); score events continue to be accepted. Always prefer upgrading to the current Python and JS SDKs. If you use custom auto-instrumentation, only then send traces via the OpenTelemetry endpoint (`POST /api/public/otel/v1/traces`), for example with curl. Learn more: https://langfuse.com/integrations/native/opentelemetry
 
         Within each batch, there can be multiple events.
         Each event has a type, an id, a timestamp, metadata and a body.

langfuse/api/ingestion/raw_client.py

@@ -27,7 +27,7 @@ def __init__(self, *, client_wrapper: SyncClientWrapper):
         self._client_wrapper = client_wrapper
 
     @typing_extensions.deprecated(
-        "On Langfuse Cloud, Langfuse v3 is deprecated and this endpoint will be removed on November 16, 2026. Send data via the OpenTelemetry endpoint at `POST /api/public/otel/v1/traces` instead, see the [OpenTelemetry integration docs](https://langfuse.com/integrations/native/opentelemetry). Self-hosted deployments are unaffected by this date; the endpoint becomes unavailable when they upgrade to Langfuse v4.",
+        "On Langfuse Cloud, Langfuse v3 is deprecated and v4-only write mode begins on November 16, 2026. This endpoint is never shut down; it continues to accept score events. Trace and observation events fail only in v4-only write mode, not in dual or legacy mode. Always prefer upgrading to the current Python and JS SDKs. If you use custom auto-instrumentation, only then send data via the OpenTelemetry endpoint at `POST /api/public/otel/v1/traces` (for example with curl); see the [OpenTelemetry integration docs](https://langfuse.com/integrations/native/opentelemetry). Self-hosted deployments are unaffected by this date; they reject trace and observation events only in v4-only write mode, not dual or legacy.",
         category=None,
     )
     def batch(
@@ -40,7 +40,7 @@ def batch(
         """
         **Legacy endpoint for batch ingestion for Langfuse Observability.**
 
-        -> Please use the OpenTelemetry endpoint (`/api/public/otel/v1/traces`). Learn more: https://langfuse.com/integrations/native/opentelemetry
+        This endpoint is never shut down. Trace and observation events are rejected only in v4-only write mode (not dual or legacy); score events continue to be accepted. Always prefer upgrading to the current Python and JS SDKs. If you use custom auto-instrumentation, only then send traces via the OpenTelemetry endpoint (`POST /api/public/otel/v1/traces`), for example with curl. Learn more: https://langfuse.com/integrations/native/opentelemetry
 
         Within each batch, there can be multiple events.
         Each event has a type, an id, a timestamp, metadata and a body.
@@ -167,7 +167,7 @@ def __init__(self, *, client_wrapper: AsyncClientWrapper):
         self._client_wrapper = client_wrapper
 
     @typing_extensions.deprecated(
-        "On Langfuse Cloud, Langfuse v3 is deprecated and this endpoint will be removed on November 16, 2026. Send data via the OpenTelemetry endpoint at `POST /api/public/otel/v1/traces` instead, see the [OpenTelemetry integration docs](https://langfuse.com/integrations/native/opentelemetry). Self-hosted deployments are unaffected by this date; the endpoint becomes unavailable when they upgrade to Langfuse v4.",
+        "On Langfuse Cloud, Langfuse v3 is deprecated and v4-only write mode begins on November 16, 2026. This endpoint is never shut down; it continues to accept score events. Trace and observation events fail only in v4-only write mode, not in dual or legacy mode. Always prefer upgrading to the current Python and JS SDKs. If you use custom auto-instrumentation, only then send data via the OpenTelemetry endpoint at `POST /api/public/otel/v1/traces` (for example with curl); see the [OpenTelemetry integration docs](https://langfuse.com/integrations/native/opentelemetry). Self-hosted deployments are unaffected by this date; they reject trace and observation events only in v4-only write mode, not dual or legacy.",
         category=None,
     )
     async def batch(
@@ -180,7 +180,7 @@ async def batch(
         """
         **Legacy endpoint for batch ingestion for Langfuse Observability.**
 
-        -> Please use the OpenTelemetry endpoint (`/api/public/otel/v1/traces`). Learn more: https://langfuse.com/integrations/native/opentelemetry
+        This endpoint is never shut down. Trace and observation events are rejected only in v4-only write mode (not dual or legacy); score events continue to be accepted. Always prefer upgrading to the current Python and JS SDKs. If you use custom auto-instrumentation, only then send traces via the OpenTelemetry endpoint (`POST /api/public/otel/v1/traces`), for example with curl. Learn more: https://langfuse.com/integrations/native/opentelemetry
 
         Within each batch, there can be multiple events.
         Each event has a type, an id, a timestamp, metadata and a body.

langfuse/api/ingestion/types/ingestion_response.py

@@ -3,14 +3,20 @@
 import typing
 
 import pydantic
+import typing_extensions
+from ...commons.types.deprecation import Deprecation
 from ...core.pydantic_utilities import UniversalBaseModel
+from ...core.serialization import FieldMetadata
 from .ingestion_error import IngestionError
 from .ingestion_success import IngestionSuccess
 
 
 class IngestionResponse(UniversalBaseModel):
     successes: typing.List[IngestionSuccess]
     errors: typing.List[IngestionError]
+    deprecation: typing_extensions.Annotated[
+        typing.Optional[Deprecation], FieldMetadata(alias="_deprecation")
+    ] = None
 
     model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(
         extra="allow", frozen=True

langfuse/api/legacy/observations_v1/client.py

@@ -35,6 +35,7 @@ def get(
         self,
         observation_id: str,
         *,
+        start_time: typing.Optional[dt.datetime] = None,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> ObservationsViewSingle:
         """
@@ -45,6 +46,9 @@ def get(
         observation_id : str
             The unique langfuse identifier of an observation, can be an event, span or generation
 
+        start_time : typing.Optional[dt.datetime]
+            The start time of the observation (ISO 8601 with offset, e.g. 2024-01-01T00:00:00Z). Optional performance hint - when provided, Langfuse narrows the lookup to make the request substantially faster. It only affects speed - an incorrect or omitted value never changes the result.
+
         request_options : typing.Optional[RequestOptions]
             Request-specific configuration.
 
@@ -69,7 +73,7 @@ def get(
         )
         """
         _response = self._raw_client.get(
-            observation_id, request_options=request_options
+            observation_id, start_time=start_time, request_options=request_options
         )
         return _response.data
 
@@ -291,6 +295,7 @@ async def get(
         self,
         observation_id: str,
         *,
+        start_time: typing.Optional[dt.datetime] = None,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> ObservationsViewSingle:
         """
@@ -301,6 +306,9 @@ async def get(
         observation_id : str
             The unique langfuse identifier of an observation, can be an event, span or generation
 
+        start_time : typing.Optional[dt.datetime]
+            The start time of the observation (ISO 8601 with offset, e.g. 2024-01-01T00:00:00Z). Optional performance hint - when provided, Langfuse narrows the lookup to make the request substantially faster. It only affects speed - an incorrect or omitted value never changes the result.
+
         request_options : typing.Optional[RequestOptions]
             Request-specific configuration.
 
@@ -333,7 +341,7 @@ async def main() -> None:
         asyncio.run(main())
         """
         _response = await self._raw_client.get(
-            observation_id, request_options=request_options
+            observation_id, start_time=start_time, request_options=request_options
         )
         return _response.data