vellum-workflow-server 1.1.1__tar.gz → 1.1.2.post1__tar.gz

{vellum_workflow_server-1.1.1 → vellum_workflow_server-1.1.2.post1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vellum-workflow-server
-Version: 1.1.1
+Version: 1.1.2.post1
 Summary: 
 License: AGPL
 Requires-Python: >=3.9.0,<4
@@ -29,7 +29,7 @@ Requires-Dist: pyjwt (==2.10.0)
 Requires-Dist: python-dotenv (==1.0.1)
 Requires-Dist: retrying (==1.3.4)
 Requires-Dist: sentry-sdk[flask] (==2.20.0)
-Requires-Dist: vellum-ai (==1.1.1)
+Requires-Dist: vellum-ai (==1.1.2)
 Description-Content-Type: text/markdown
 
 # Vellum Workflow Runner Server

{vellum_workflow_server-1.1.1 → vellum_workflow_server-1.1.2.post1}/pyproject.toml

@@ -3,7 +3,7 @@ name = "vellum-workflow-server"
 
 [tool.poetry]
 name = "vellum-workflow-server"
-version = "1.1.1"
+version = "1.1.2.post1"
 description = ""
 readme = "README.md"
 authors = []
@@ -45,7 +45,7 @@ flask = "2.3.3"
 orderly-set = "5.2.2"
 pebble = "5.0.7"
 gunicorn = "23.0.0"
-vellum-ai = "1.1.1"
+vellum-ai = "1.1.2"
 python-dotenv = "1.0.1"
 retrying = "1.3.4"
 sentry-sdk = {extras = ["flask"], version = "2.20.0"}

{vellum_workflow_server-1.1.1 → vellum_workflow_server-1.1.2.post1}/src/workflow_server/api/workflow_view.py

@@ -61,6 +61,17 @@ CUSTOM_NODES_DIRECTORY = "vellum_custom_nodes"
 WORKFLOW_INITIATION_TIMEOUT_SECONDS = 60
 
 
+def _events_emission_enabled(context: Union[WorkflowExecutorContext, NodeExecutorContext]) -> bool:
+    flags = context.feature_flags
+    if not (flags and flags.get("vembda-event-emitting-enabled")):
+        return False
+    try:
+        module = importlib.import_module("vellum.workflows.events.emitters.vellum")
+        return hasattr(module, "VellumEmitter")
+    except Exception:
+        return False
+
+
 @bp.route("/stream", methods=["POST"])
 def stream_workflow_route() -> Response:
     data = request.get_json()
@@ -97,6 +108,7 @@ def stream_workflow_route() -> Response:
     headers = {
         "X-Vellum-SDK-Version": vembda_initiated_event.body.sdk_version,
         "X-Vellum-Server-Version": vembda_initiated_event.body.server_version,
+        "X-Vellum-Events-Emitted": str(_events_emission_enabled(context)),
     }
 
     # We can exceed the concurrency count currently with long running workflows due to a knative issue. So here
@@ -403,6 +415,7 @@ def stream_node_route() -> Response:
     headers = {
         "X-Vellum-SDK-Version": vembda_initiated_event.body.sdk_version,
         "X-Vellum-Server-Version": vembda_initiated_event.body.server_version,
+        "X-Vellum-Events-Emitted": str(_events_emission_enabled(context)),
     }
 
     resp = Response(

{vellum_workflow_server-1.1.1 → vellum_workflow_server-1.1.2.post1}/src/workflow_server/core/executor.py

@@ -12,7 +12,7 @@ from threading import Event as ThreadingEvent
 import time
 from traceback import format_exc
 from uuid import UUID, uuid4
-from typing import Any, Callable, Generator, Iterator, Optional, Tuple, Type
+from typing import TYPE_CHECKING, Any, Callable, Generator, Iterator, Optional, Tuple, Type
 
 from pebble import concurrent
 from vellum_ee.workflows.display.workflows import BaseWorkflowDisplay
@@ -47,6 +47,15 @@ from workflow_server.core.workflow_executor_context import (
 )
 from workflow_server.utils.log_proxy import redirect_log
 
+if TYPE_CHECKING:
+    # Imported for typing only to avoid runtime dependency issues
+    from vellum.workflows.events.emitters.base import BaseWorkflowEmitter
+
+try:
+    from vellum.workflows.events.emitters.vellum import VellumEmitter as _VellumEmitter
+except Exception:
+    _VellumEmitter = None
+
 logger = logging.getLogger(__name__)
 
 
@@ -331,13 +340,35 @@ def _create_workflow(executor_context: WorkflowExecutorContext, namespace: str)
     VembdaExecutionFulfilledEvent.model_rebuild(
         # Not sure why this is needed, but it is required for the VembdaExecutionFulfilledEvent to be
         # properly rebuilt with the recursive types.
+        # use flag here to determine which emitter to use
         _types_namespace={
             "BaseWorkflow": BaseWorkflow,
             "BaseNode": BaseNode,
         },
     )
 
-    return Workflow(context=workflow_context, store=EmptyStore())
+    # Determine whether to enable the Vellum Emitter for event publishing
+    use_vellum_emitter = bool((executor_context.feature_flags or {}).get("vembda-event-emitting-enabled"))
+
+    emitters: list["BaseWorkflowEmitter"] = []
+    if use_vellum_emitter:
+        if _VellumEmitter is not None:
+            try:
+                emitters = [_VellumEmitter()]
+            except Exception:
+                emitters = []
+        else:
+            # Emitter is not available at import time; continue without it
+            logger.warning("VellumEmitter not available; continuing without it")
+
+    # Explicit constructor call to satisfy typing
+    workflow = Workflow(
+        context=workflow_context,
+        store=EmptyStore(),
+        emitters=emitters or None,
+    )
+
+    return workflow
 
 
 def _create_workflow_context(executor_context: BaseExecutorContext) -> WorkflowContext:

{vellum_workflow_server-1.1.1 → vellum_workflow_server-1.1.2.post1}/src/workflow_server/core/workflow_executor_context.py

@@ -24,6 +24,7 @@ class BaseExecutorContext(UniversalBaseModel):
     vembda_public_url: Optional[str] = None
     node_output_mocks: Optional[list[Any]] = None
     environment_variables: Optional[dict[str, str]] = None
+    feature_flags: Optional[dict[str, bool]] = None
 
     @property
     def container_overhead_latency(self) -> int: