agno 2.0.11__py3-none-any.whl → 2.1.1__py3-none-any.whl

agno/vectordb/upstashdb/upstashdb.py

@@ -504,8 +504,48 @@ class UpstashVectorDb(VectorDb):
         _namespace = self.namespace if namespace is None else namespace
         vectors = []
 
-        embed_tasks = [document.async_embed(embedder=self.embedder) for document in documents]
-        await asyncio.gather(*embed_tasks, return_exceptions=True)
+        if (
+            self.embedder
+            and self.embedder.enable_batch
+            and hasattr(self.embedder, "async_get_embeddings_batch_and_usage")
+        ):
+            # Use batch embedding when enabled and supported
+            try:
+                # Extract content from all documents
+                doc_contents = [doc.content for doc in documents]
+
+                # Get batch embeddings and usage
+                embeddings, usages = await self.embedder.async_get_embeddings_batch_and_usage(doc_contents)
+
+                # Process documents with pre-computed embeddings
+                for j, doc in enumerate(documents):
+                    try:
+                        if j < len(embeddings):
+                            doc.embedding = embeddings[j]
+                            doc.usage = usages[j] if j < len(usages) else None
+                    except Exception as e:
+                        logger.error(f"Error assigning batch embedding to document '{doc.name}': {e}")
+
+            except Exception as e:
+                # Check if this is a rate limit error - don't fall back as it would make things worse
+                error_str = str(e).lower()
+                is_rate_limit = any(
+                    phrase in error_str
+                    for phrase in ["rate limit", "too many requests", "429", "trial key", "api calls / minute"]
+                )
+
+                if is_rate_limit:
+                    logger.error(f"Rate limit detected during batch embedding. {e}")
+                    raise e
+                else:
+                    logger.warning(f"Async batch embedding failed, falling back to individual embeddings: {e}")
+                    # Fall back to individual embedding
+                    embed_tasks = [doc.async_embed(embedder=self.embedder) for doc in documents]
+                    await asyncio.gather(*embed_tasks, return_exceptions=True)
+        else:
+            # Use individual embedding
+            embed_tasks = [document.async_embed(embedder=self.embedder) for document in documents]
+            await asyncio.gather(*embed_tasks, return_exceptions=True)
 
         for i, document in enumerate(documents):
             if document.id is None:

agno/vectordb/weaviate/weaviate.py

@@ -270,9 +270,45 @@ class Weaviate(VectorDb):
         if not documents:
             return
 
-        # Embed document
-        embed_tasks = [document.async_embed(embedder=self.embedder) for document in documents]
-        await asyncio.gather(*embed_tasks, return_exceptions=True)
+        # Apply batch embedding logic
+        if self.embedder.enable_batch and hasattr(self.embedder, "async_get_embeddings_batch_and_usage"):
+            # Use batch embedding when enabled and supported
+            try:
+                # Extract content from all documents
+                doc_contents = [doc.content for doc in documents]
+
+                # Get batch embeddings and usage
+                embeddings, usages = await self.embedder.async_get_embeddings_batch_and_usage(doc_contents)
+
+                # Process documents with pre-computed embeddings
+                for j, doc in enumerate(documents):
+                    try:
+                        if j < len(embeddings):
+                            doc.embedding = embeddings[j]
+                            doc.usage = usages[j] if j < len(usages) else None
+                    except Exception as e:
+                        logger.error(f"Error assigning batch embedding to document '{doc.name}': {e}")
+
+            except Exception as e:
+                # Check if this is a rate limit error - don't fall back as it would make things worse
+                error_str = str(e).lower()
+                is_rate_limit = any(
+                    phrase in error_str
+                    for phrase in ["rate limit", "too many requests", "429", "trial key", "api calls / minute"]
+                )
+
+                if is_rate_limit:
+                    logger.error(f"Rate limit detected during batch embedding. {e}")
+                    raise e
+                else:
+                    logger.warning(f"Async batch embedding failed, falling back to individual embeddings: {e}")
+                    # Fall back to individual embedding
+                    embed_tasks = [doc.async_embed(embedder=self.embedder) for doc in documents]
+                    await asyncio.gather(*embed_tasks, return_exceptions=True)
+        else:
+            # Use individual embedding
+            embed_tasks = [document.async_embed(embedder=self.embedder) for document in documents]
+            await asyncio.gather(*embed_tasks, return_exceptions=True)
 
         client = await self.get_async_client()
         try:

agno/workflow/types.py

@@ -1,3 +1,4 @@
+import json
 from dataclasses import dataclass
 from enum import Enum
 from typing import Any, Dict, List, Optional, Union
@@ -8,6 +9,7 @@ from pydantic import BaseModel
 from agno.media import Audio, File, Image, Video
 from agno.models.metrics import Metrics
 from agno.utils.log import log_warning
+from agno.utils.serialize import json_serializer
 
 
 @dataclass
@@ -57,6 +59,7 @@ class WorkflowExecutionInput:
             "images": [img.to_dict() for img in self.images] if self.images else None,
             "videos": [vid.to_dict() for vid in self.videos] if self.videos else None,
             "audio": [aud.to_dict() for aud in self.audio] if self.audio else None,
+            "files": [file.to_dict() for file in self.files] if self.files else None,
         }
 
 
@@ -442,9 +445,7 @@ class WebSocketHandler:
             else:
                 data = {"type": "message", "content": str(event)}
 
-            import json
-
-            await self.websocket.send_text(self.format_sse_event(json.dumps(data)))
+            await self.websocket.send_text(self.format_sse_event(json.dumps(data, default=json_serializer)))
 
         except Exception as e:
             log_warning(f"Failed to handle WebSocket event: {e}")
@@ -465,9 +466,7 @@ class WebSocketHandler:
             return
 
         try:
-            import json
-
-            await self.websocket.send_text(self.format_sse_event(json.dumps(data)))
+            await self.websocket.send_text(self.format_sse_event(json.dumps(data, default=json_serializer)))
         except Exception as e:
             log_warning(f"Failed to send WebSocket dict: {e}")
 

agno/workflow/workflow.py

@@ -25,7 +25,7 @@ from pydantic import BaseModel
 
 from agno.agent.agent import Agent
 from agno.db.base import BaseDb, SessionType
-from agno.exceptions import RunCancelledException
+from agno.exceptions import InputCheckError, OutputCheckError, RunCancelledException
 from agno.media import Audio, File, Image, Video
 from agno.models.message import Message
 from agno.models.metrics import Metrics
@@ -54,6 +54,7 @@ from agno.team.team import Team
 from agno.utils.common import is_typed_dict, validate_typed_dict
 from agno.utils.log import (
     log_debug,
+    log_error,
     log_warning,
     logger,
     set_log_level_to_debug,
@@ -1000,8 +1001,14 @@ class Workflow:
                 workflow_run_response.audio = output_audio
                 workflow_run_response.status = RunStatus.completed
 
+            except (InputCheckError, OutputCheckError) as e:
+                log_error(f"Validation failed: {str(e)} | Check: {e.check_trigger}")
+                # Store error response
+                workflow_run_response.status = RunStatus.error
+                workflow_run_response.content = f"Validation failed: {str(e)} | Check: {e.check_trigger}"
+
+                raise e
             except RunCancelledException as e:
-                # Handle run cancellation
                 logger.info(f"Workflow run {workflow_run_response.run_id} was cancelled")
                 workflow_run_response.status = RunStatus.cancelled
                 workflow_run_response.content = str(e)
@@ -1013,6 +1020,7 @@ class Workflow:
                 # Store error response
                 workflow_run_response.status = RunStatus.error
                 workflow_run_response.content = f"Workflow execution failed: {e}"
+                raise e
 
             finally:
                 self._update_session_metrics(session=session, workflow_run_response=workflow_run_response)
@@ -1199,6 +1207,24 @@ class Workflow:
                 workflow_run_response.audio = output_audio
                 workflow_run_response.status = RunStatus.completed
 
+            except (InputCheckError, OutputCheckError) as e:
+                log_error(f"Validation failed: {str(e)} | Check: {e.check_trigger}")
+
+                from agno.run.workflow import WorkflowErrorEvent
+
+                error_event = WorkflowErrorEvent(
+                    run_id=workflow_run_response.run_id or "",
+                    workflow_id=self.id,
+                    workflow_name=self.name,
+                    session_id=session.session_id,
+                    error=str(e),
+                )
+
+                yield error_event
+
+                # Update workflow_run_response with error
+                workflow_run_response.content = error_event.error
+                workflow_run_response.status = RunStatus.error
             except RunCancelledException as e:
                 # Handle run cancellation during streaming
                 logger.info(f"Workflow run {workflow_run_response.run_id} was cancelled during streaming")
@@ -1230,6 +1256,7 @@ class Workflow:
                 # Update workflow_run_response with error
                 workflow_run_response.content = error_event.error
                 workflow_run_response.status = RunStatus.error
+                raise e
 
         # Yield workflow completed event
         workflow_completed_event = WorkflowCompletedEvent(
@@ -1436,6 +1463,13 @@ class Workflow:
                 workflow_run_response.audio = output_audio
                 workflow_run_response.status = RunStatus.completed
 
+            except (InputCheckError, OutputCheckError) as e:
+                log_error(f"Validation failed: {str(e)} | Check: {e.check_trigger}")
+                # Store error response
+                workflow_run_response.status = RunStatus.error
+                workflow_run_response.content = f"Validation failed: {str(e)} | Check: {e.check_trigger}"
+
+                raise e
             except RunCancelledException as e:
                 logger.info(f"Workflow run {workflow_run_response.run_id} was cancelled")
                 workflow_run_response.status = RunStatus.cancelled
@@ -1444,6 +1478,7 @@ class Workflow:
                 logger.error(f"Workflow execution failed: {e}")
                 workflow_run_response.status = RunStatus.error
                 workflow_run_response.content = f"Workflow execution failed: {e}"
+                raise e
 
         self._update_session_metrics(session=session, workflow_run_response=workflow_run_response)
         session.upsert_run(run=workflow_run_response)
@@ -1471,6 +1506,7 @@ class Workflow:
         from inspect import isasyncgenfunction, iscoroutinefunction, isgeneratorfunction
 
         workflow_run_response.status = RunStatus.running
+
         workflow_started_event = WorkflowStartedEvent(
             run_id=workflow_run_response.run_id or "",
             workflow_name=workflow_run_response.workflow_name,
@@ -1633,6 +1669,24 @@ class Workflow:
                 workflow_run_response.audio = output_audio
                 workflow_run_response.status = RunStatus.completed
 
+            except (InputCheckError, OutputCheckError) as e:
+                log_error(f"Validation failed: {str(e)} | Check: {e.check_trigger}")
+
+                from agno.run.workflow import WorkflowErrorEvent
+
+                error_event = WorkflowErrorEvent(
+                    run_id=workflow_run_response.run_id or "",
+                    workflow_id=self.id,
+                    workflow_name=self.name,
+                    session_id=session.session_id,
+                    error=str(e),
+                )
+
+                yield error_event
+
+                # Update workflow_run_response with error
+                workflow_run_response.content = error_event.error
+                workflow_run_response.status = RunStatus.error
             except RunCancelledException as e:
                 # Handle run cancellation during streaming
                 logger.info(f"Workflow run {workflow_run_response.run_id} was cancelled during streaming")
@@ -1668,6 +1722,7 @@ class Workflow:
                 # Update workflow_run_response with error
                 workflow_run_response.content = error_event.error
                 workflow_run_response.status = RunStatus.error
+                raise e
 
         # Yield workflow completed event
         workflow_completed_event = WorkflowCompletedEvent(
@@ -2175,6 +2230,7 @@ class Workflow:
             audio=audio,  # type: ignore
             images=images,  # type: ignore
             videos=videos,  # type: ignore
+            files=files,
         )
         log_debug(
             f"Created async pipeline input with session state keys: {list(session_state.keys()) if session_state else 'None'}"

{agno-2.0.11.dist-info → agno-2.1.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agno
-Version: 2.0.11
+Version: 2.1.1
 Summary: Agno: a lightweight library for building Multi-Agent Systems
 Author-email: Ashpreet Bedi <ashpreet@agno.com>
 Project-URL: homepage, https://agno.com
@@ -48,6 +48,7 @@ Requires-Dist: uvicorn; extra == "dev"
 Provides-Extra: os
 Requires-Dist: fastapi; extra == "os"
 Requires-Dist: uvicorn; extra == "os"
+Requires-Dist: PyJWT; extra == "os"
 Provides-Extra: integration-tests
 Requires-Dist: exa_py; extra == "integration-tests"
 Requires-Dist: ddgs; extra == "integration-tests"
@@ -444,7 +445,7 @@ At Agno, we're obsessed with performance. Why? because even simple AI workflows
 - Agent instantiation: ~3μs on average
 - Memory footprint: ~6.5Kib on average
 
-> Tested on an Apple M4 Mackbook Pro.
+> Tested on an Apple M4 MacBook Pro.
 
 While an Agent's run-time is bottlenecked by inference, we must do everything possible to minimize execution time, reduce memory usage, and parallelize tool calls. These numbers may seem trivial at first, but our experience shows that they add up even at a reasonably small scale.
 
@@ -468,7 +469,7 @@ python evals/performance/instantiation_with_tool.py
 python evals/performance/other/langgraph_instantiation.py
 ```
 
-> The following evaluation is run on an Apple M4 Mackbook Pro. It also runs as a Github action on this repo.
+> The following evaluation is run on an Apple M4 MacBook Pro. It also runs as a Github action on this repo.
 
 LangGraph is on the right, **let's start it first and give it a head start**.