agno 2.0.5__py3-none-any.whl → 2.0.6__py3-none-any.whl

agno/agent/agent.py

@@ -371,7 +371,7 @@ class Agent:
         knowledge_retriever: Optional[Callable[..., Optional[List[Union[Dict, str]]]]] = None,
         references_format: Literal["json", "yaml"] = "json",
         metadata: Optional[Dict[str, Any]] = None,
-        tools: Optional[List[Union[Toolkit, Callable, Function, Dict]]] = None,
+        tools: Optional[Sequence[Union[Toolkit, Callable, Function, Dict]]] = None,
         tool_call_limit: Optional[int] = None,
         tool_choice: Optional[Union[str, Dict[str, Any]]] = None,
         tool_hooks: Optional[List[Callable]] = None,
@@ -466,7 +466,7 @@ class Agent:
 
         self.metadata = metadata
 
-        self.tools = tools
+        self.tools = list(tools) if tools else []
         self.tool_call_limit = tool_call_limit
         self.tool_choice = tool_choice
         self.tool_hooks = tool_hooks
@@ -681,8 +681,8 @@ class Agent:
         self.tools.append(tool)
         self._rebuild_tools = True
 
-    def set_tools(self, tools: List[Union[Toolkit, Callable, Function, Dict]]):
-        self.tools = tools
+    def set_tools(self, tools: Sequence[Union[Toolkit, Callable, Function, Dict]]):
+        self.tools = list(tools) if tools else []
         self._rebuild_tools = True
 
     def _initialize_session(
@@ -1080,13 +1080,17 @@ class Agent:
         # Initialize the Agent
         self.initialize_agent(debug_mode=debug_mode)
 
-        image_artifacts, video_artifacts, audio_artifacts = self._validate_media_object_id(
-            images=images, videos=videos, audios=audio
+        image_artifacts, video_artifacts, audio_artifacts, file_artifacts = self._validate_media_object_id(
+            images=images, videos=videos, audios=audio, files=files
         )
 
         # Create RunInput to capture the original user input
         run_input = RunInput(
-            input_content=input, images=image_artifacts, videos=video_artifacts, audios=audio_artifacts, files=files
+            input_content=input,
+            images=image_artifacts,
+            videos=video_artifacts,
+            audios=audio_artifacts,
+            files=file_artifacts,
         )
 
         # Read existing session from database
@@ -1713,13 +1717,17 @@ class Agent:
         # Initialize the Agent
         self.initialize_agent(debug_mode=debug_mode)
 
-        image_artifacts, video_artifacts, audio_artifacts = self._validate_media_object_id(
-            images=images, videos=videos, audios=audio
+        image_artifacts, video_artifacts, audio_artifacts, file_artifacts = self._validate_media_object_id(
+            images=images, videos=videos, audios=audio, files=files
         )
 
         # Create RunInput to capture the original user input
         run_input = RunInput(
-            input_content=input, images=image_artifacts, videos=video_artifacts, audios=audio_artifacts, files=files
+            input_content=input,
+            images=image_artifacts,
+            videos=video_artifacts,
+            audios=audio_artifacts,
+            files=file_artifacts,
         )
 
         # Read existing session from storage
@@ -3063,6 +3071,10 @@ class Agent:
             for audio in model_response.audios:
                 self._add_audio(audio, run_response)  # Generated audio go to run_response.audio
 
+        if model_response.files is not None:
+            for file in model_response.files:
+                self._add_file(file, run_response)  # Generated files go to run_response.files
+
     def _update_run_response(self, model_response: ModelResponse, run_response: RunOutput, run_messages: RunMessages):
         # Handle structured outputs
         if self.output_schema is not None and model_response.parsed is not None:
@@ -3125,7 +3137,8 @@ class Agent:
         """Calculate session metrics"""
         session_metrics = self._get_session_metrics(session=session)
         # Add the metrics for the current run to the session metrics
-        session_metrics += run_response.metrics
+        if run_response.metrics is not None:
+            session_metrics += run_response.metrics
         session_metrics.time_to_first_token = None
         if session.session_data is not None:
             session.session_data["session_metrics"] = session_metrics
@@ -3800,7 +3813,7 @@ class Agent:
             self._rebuild_tools = True
         if self.search_session_history:
             agent_tools.append(
-                self._get_previous_sessions_messages_function(num_history_sessions=self.num_history_sessions)
+                self._get_previous_sessions_messages_function(num_history_sessions=self.num_history_sessions, user_id=user_id)
             )
             self._rebuild_tools = True
 
@@ -4948,7 +4961,7 @@ class Agent:
         # 1. If build_user_context is False or message is a list, return the message as is.
         if not self.build_user_context:
             return Message(
-                role=self.user_message_role,
+                role=self.user_message_role or "user",
                 content=input,
                 images=None if not self.send_media_to_model else images,
                 audio=None if not self.send_media_to_model else audio,
@@ -4961,7 +4974,7 @@ class Agent:
             # If we have any media, return a message with empty content
             if images is not None or audio is not None or videos is not None or files is not None:
                 return Message(
-                    role=self.user_message_role,
+                    role=self.user_message_role or "user",
                     content="",
                     images=None if not self.send_media_to_model else images,
                     audio=None if not self.send_media_to_model else audio,
@@ -5716,6 +5729,13 @@ class Agent:
             run_response.audio = []
         run_response.audio.append(audio)
 
+    def _add_file(self, file: File, run_response: RunOutput) -> None:
+        """Add file to both the agent's stateful storage and the current run response"""
+        # Add to run response
+        if run_response.files is None:
+            run_response.files = []
+        run_response.files.append(file)
+
     ###########################################################################
     # Reasoning
     ###########################################################################
@@ -6779,11 +6799,14 @@ class Agent:
         )
         return "Successfully added to knowledge base"
 
-    def _get_previous_sessions_messages_function(self, num_history_sessions: Optional[int] = 2) -> Callable:
+    def _get_previous_sessions_messages_function(
+        self, num_history_sessions: Optional[int] = 2, user_id: Optional[str] = None
+    ) -> Callable:
         """Factory function to create a get_previous_session_messages function.
 
         Args:
             num_history_sessions: The last n sessions to be taken from db
+            user_id: The user ID to filter sessions by
 
         Returns:
             Callable: A function that retrieves messages from previous sessions
@@ -6802,7 +6825,9 @@ class Agent:
             if self.db is None:
                 return "Previous session messages not available"
 
-            selected_sessions = self.db.get_sessions(session_type=SessionType.AGENT, limit=num_history_sessions)
+            selected_sessions = self.db.get_sessions(
+                session_type=SessionType.AGENT, limit=num_history_sessions, user_id=user_id
+            )
 
             all_messages = []
             seen_message_pairs = set()
@@ -7199,6 +7224,7 @@ class Agent:
         images: Optional[Sequence[Image]] = None,
         videos: Optional[Sequence[Video]] = None,
         audios: Optional[Sequence[Audio]] = None,
+        files: Optional[Sequence[File]] = None,
     ) -> tuple:
         """Convert raw Image/Video/Audio objects - now unified, so just return as-is."""
         # With unified classes, no conversion needed - just ensure IDs are set
@@ -7233,7 +7259,17 @@ class Agent:
                     aud.id = str(uuid4())
                 audio_list.append(aud)
 
-        return image_list, video_list, audio_list
+        file_list = None
+        if files:
+            file_list = []
+            for file in files:
+                if not file.id:
+                    from uuid import uuid4
+
+                    file.id = str(uuid4())
+                file_list.append(file)
+
+        return image_list, video_list, audio_list, file_list
 
     def cli_app(
         self,

agno/db/dynamo/dynamo.py

@@ -181,7 +181,7 @@ class DynamoDb(BaseDb):
 
     # --- Sessions ---
 
-    def delete_session(self, session_id: Optional[str] = None, session_type: Optional[SessionType] = None) -> bool:
+    def delete_session(self, session_id: Optional[str] = None) -> bool:
         """
         Delete a session from the database.
 
@@ -236,7 +236,7 @@ class DynamoDb(BaseDb):
     def get_session(
         self,
         session_id: str,
-        session_type: Optional[SessionType] = None,
+        session_type: SessionType,
         user_id: Optional[str] = None,
         deserialize: Optional[bool] = True,
     ) -> Optional[Union[Session, Dict[str, Any]]]:
@@ -245,7 +245,7 @@ class DynamoDb(BaseDb):
 
         Args:
             session_id (str): The ID of the session to get.
-            session_type (Optional[SessionType]): The type of session to get.
+            session_type (SessionType): The type of session to get.
             user_id (Optional[str]): The ID of the user to get the session for.
             deserialize (Optional[bool]): Whether to deserialize the session.
 
@@ -268,7 +268,7 @@ class DynamoDb(BaseDb):
 
             session = deserialize_from_dynamodb_item(item)
 
-            if session_type and session.get("session_type") != session_type.value:
+            if session.get("session_type") != session_type.value:
                 return None
             if user_id and session.get("user_id") != user_id:
                 return None
@@ -283,8 +283,10 @@ class DynamoDb(BaseDb):
                 return AgentSession.from_dict(session)
             elif session_type == SessionType.TEAM:
                 return TeamSession.from_dict(session)
-            else:
+            elif session_type == SessionType.WORKFLOW:
                 return WorkflowSession.from_dict(session)
+            else:
+                raise ValueError(f"Invalid session type: {session_type}")
 
         except Exception as e:
             log_error(f"Failed to get session {session_id}: {e}")

agno/db/firestore/firestore.py

@@ -242,8 +242,8 @@ class FirestoreDb(BaseDb):
 
         Args:
             session_id (str): The ID of the session to get.
+            session_type (SessionType): The type of session to get.
             user_id (Optional[str]): The ID of the user to get the session for.
-            session_type (Optional[SessionType]): The type of session to get.
             deserialize (Optional[bool]): Whether to serialize the session. Defaults to True.
 
         Returns:
@@ -281,8 +281,10 @@ class FirestoreDb(BaseDb):
                 return AgentSession.from_dict(session)
             elif session_type == SessionType.TEAM:
                 return TeamSession.from_dict(session)
-            else:
+            elif session_type == SessionType.WORKFLOW:
                 return WorkflowSession.from_dict(session)
+            else:
+                raise ValueError(f"Invalid session type: {session_type}")
 
         except Exception as e:
             log_error(f"Exception reading session: {e}")

agno/db/gcs_json/gcs_json_db.py

@@ -185,7 +185,7 @@ class GcsJsonDb(BaseDb):
     def get_session(
         self,
         session_id: str,
-        session_type: Optional[SessionType] = None,
+        session_type: SessionType,
         user_id: Optional[str] = None,
         deserialize: Optional[bool] = True,
     ) -> Optional[Union[AgentSession, TeamSession, WorkflowSession, Dict[str, Any]]]:
@@ -193,7 +193,7 @@ class GcsJsonDb(BaseDb):
 
         Args:
             session_id (str): The ID of the session to read.
-            session_type (Optional[SessionType]): The type of the session to read.
+            session_type (SessionType): The type of the session to read.
             user_id (Optional[str]): The ID of the user to read the session for.
             deserialize (Optional[bool]): Whether to deserialize the session.
 
@@ -226,6 +226,8 @@ class GcsJsonDb(BaseDb):
                         return TeamSession.from_dict(session_data)
                     elif session_type == SessionType.WORKFLOW:
                         return WorkflowSession.from_dict(session_data)
+                    else:
+                        raise ValueError(f"Invalid session type: {session_type}")
 
             return None
 

agno/db/json/json_db.py

@@ -168,7 +168,7 @@ class JsonDb(BaseDb):
     def get_session(
         self,
         session_id: str,
-        session_type: Optional[SessionType] = None,
+        session_type: SessionType,
         user_id: Optional[str] = None,
         deserialize: Optional[bool] = True,
     ) -> Optional[Union[AgentSession, TeamSession, WorkflowSession, Dict[str, Any]]]:
@@ -176,7 +176,7 @@ class JsonDb(BaseDb):
 
         Args:
             session_id (str): The ID of the session to read.
-            session_type (Optional[SessionType]): The type of the session to read.
+            session_type (SessionType): The type of the session to read.
             user_id (Optional[str]): The ID of the user to read the session for.
             deserialize (Optional[bool]): Whether to deserialize the session.
 
@@ -208,8 +208,10 @@ class JsonDb(BaseDb):
                         return AgentSession.from_dict(session)
                     elif session_type == SessionType.TEAM:
                         return TeamSession.from_dict(session)
-                    else:
+                    elif session_type == SessionType.WORKFLOW:
                         return WorkflowSession.from_dict(session)
+                    else:
+                        raise ValueError(f"Invalid session type: {session_type}")
 
             return None
 
@@ -338,8 +340,10 @@ class JsonDb(BaseDb):
                         return AgentSession.from_dict(session)
                     elif session_type == SessionType.TEAM:
                         return TeamSession.from_dict(session)
-                    else:
+                    elif session_type == SessionType.WORKFLOW:
                         return WorkflowSession.from_dict(session)
+                    else:
+                        raise ValueError(f"Invalid session type: {session_type}")
 
             return None
 

agno/db/mongo/mongo.py

@@ -253,8 +253,8 @@ class MongoDb(BaseDb):
 
         Args:
             session_id (str): The ID of the session to get.
+            session_type (SessionType): The type of session to get.
             user_id (Optional[str]): The ID of the user to get the session for.
-            session_type (Optional[SessionType]): The type of session to get.
             deserialize (Optional[bool]): Whether to serialize the session. Defaults to True.
 
         Returns:
@@ -285,12 +285,14 @@ class MongoDb(BaseDb):
             if not deserialize:
                 return session
 
-            if session_type == SessionType.AGENT.value:
+            if session_type == SessionType.AGENT:
                 return AgentSession.from_dict(session)
-            elif session_type == SessionType.TEAM.value:
+            elif session_type == SessionType.TEAM:
                 return TeamSession.from_dict(session)
-            else:
+            elif session_type == SessionType.WORKFLOW:
                 return WorkflowSession.from_dict(session)
+            else:
+                raise ValueError(f"Invalid session type: {session_type}")
 
         except Exception as e:
             log_error(f"Exception reading session: {e}")

agno/db/mysql/mysql.py

@@ -350,8 +350,8 @@ class MySQLDb(BaseDb):
 
         Args:
             session_id (str): ID of the session to read.
+            session_type (SessionType): Type of session to get.
             user_id (Optional[str]): User ID to filter by. Defaults to None.
-            session_type (Optional[SessionType]): Type of session to read. Defaults to None.
             deserialize (Optional[bool]): Whether to serialize the session. Defaults to True.
 
         Returns:
@@ -415,6 +415,7 @@ class MySQLDb(BaseDb):
         Get all sessions in the given table. Can filter by user_id and entity_id.
 
         Args:
+            session_type (Optional[SessionType]): The type of sessions to get.
             user_id (Optional[str]): The ID of the user to filter by.
             entity_id (Optional[str]): The ID of the agent / workflow to filter by.
             start_timestamp (Optional[int]): The start timestamp to filter by.

agno/db/postgres/postgres.py

@@ -345,8 +345,8 @@ class PostgresDb(BaseDb):
 
         Args:
             session_id (str): ID of the session to read.
+            session_type (SessionType): Type of session to get.
             user_id (Optional[str]): User ID to filter by. Defaults to None.
-            session_type (Optional[SessionType]): Type of session to read. Defaults to None.
             deserialize (Optional[bool]): Whether to serialize the session. Defaults to True.
 
         Returns:
@@ -410,6 +410,7 @@ class PostgresDb(BaseDb):
         Get all sessions in the given table. Can filter by user_id and entity_id.
 
         Args:
+            session_type (Optional[SessionType]): The type of session to get.
             user_id (Optional[str]): The ID of the user to filter by.
             entity_id (Optional[str]): The ID of the agent / workflow to filter by.
             start_timestamp (Optional[int]): The start timestamp to filter by.

agno/db/redis/redis.py

@@ -300,8 +300,8 @@ class RedisDb(BaseDb):
 
         Args:
             session_id (str): The ID of the session to get.
+            session_type (SessionType): The type of session to get.
             user_id (Optional[str]): The ID of the user to filter by.
-            session_type (Optional[SessionType]): The type of session to filter by.
 
         Returns:
             Optional[Union[AgentSession, TeamSession, WorkflowSession]]: The session if found, None otherwise.

agno/db/singlestore/singlestore.py

@@ -431,8 +431,8 @@ class SingleStoreDb(BaseDb):
 
         Args:
             session_id (str): ID of the session to read.
+            session_type (SessionType): Type of session to get.
             user_id (Optional[str]): User ID to filter by. Defaults to None.
-            session_type (Optional[SessionType]): Type of session to read. Defaults to None.
             deserialize (Optional[bool]): Whether to serialize the session. Defaults to True.
 
         Returns:
@@ -496,7 +496,7 @@ class SingleStoreDb(BaseDb):
         Get all sessions in the given table. Can filter by user_id and entity_id.
 
         Args:
-            session_type (Optional[SessionType]): The type of session to filter by. Defaults to None.
+            session_type (Optional[SessionType]): The type of session to filter by.
             user_id (Optional[str]): The ID of the user to filter by.
             component_id (Optional[str]): The ID of the agent / workflow to filter by.
             session_name (Optional[str]): The name of the session to filter by.

agno/db/sqlite/sqlite.py

@@ -333,8 +333,8 @@ class SqliteDb(BaseDb):
 
         Args:
             session_id (str): ID of the session to read.
+            session_type (SessionType): Type of session to get.
             user_id (Optional[str]): User ID to filter by. Defaults to None.
-            session_type (Optional[SessionType]): Type of session to read. Defaults to None.
             deserialize (Optional[bool]): Whether to serialize the session. Defaults to True.
 
         Returns:

agno/knowledge/embedder/openai.py

@@ -78,21 +78,25 @@ class OpenAIEmbedder(Embedder):
         return self.client.embeddings.create(**_request_params)
 
     def get_embedding(self, text: str) -> List[float]:
-        response: CreateEmbeddingResponse = self.response(text=text)
         try:
+            response: CreateEmbeddingResponse = self.response(text=text)
             return response.data[0].embedding
         except Exception as e:
             logger.warning(e)
             return []
 
     def get_embedding_and_usage(self, text: str) -> Tuple[List[float], Optional[Dict]]:
-        response: CreateEmbeddingResponse = self.response(text=text)
+        try:
+            response: CreateEmbeddingResponse = self.response(text=text)
 
-        embedding = response.data[0].embedding
-        usage = response.usage
-        if usage:
-            return embedding, usage.model_dump()
-        return embedding, None
+            embedding = response.data[0].embedding
+            usage = response.usage
+            if usage:
+                return embedding, usage.model_dump()
+            return embedding, None
+        except Exception as e:
+            logger.warning(e)
+            return [], None
 
     async def async_get_embedding(self, text: str) -> List[float]:
         req: Dict[str, Any] = {
@@ -127,10 +131,14 @@ class OpenAIEmbedder(Embedder):
         if self.request_params:
             req.update(self.request_params)
 
-        response = await self.aclient.embeddings.create(**req)
-        embedding = response.data[0].embedding
-        usage = response.usage
-        return embedding, usage.model_dump() if usage else None
+        try:
+            response = await self.aclient.embeddings.create(**req)
+            embedding = response.data[0].embedding
+            usage = response.usage
+            return embedding, usage.model_dump() if usage else None
+        except Exception as e:
+            logger.warning(e)
+            return [], None
 
     def get_embeddings_batch(self, texts: List[str], batch_size: int = 100) -> List[List[float]]:
         """

agno/knowledge/knowledge.py

@@ -74,6 +74,8 @@ class Knowledge:
     async def add_contents_async(self, *args, **kwargs) -> None:
         if args and isinstance(args[0], list):
             arguments = args[0]
+            upsert = kwargs.get("upsert", False)
+            skip_if_exists = kwargs.get("skip_if_exists", False)
             for argument in arguments:
                 await self.add_content_async(
                     name=argument.get("name"),
@@ -85,8 +87,8 @@ class Knowledge:
                     reader=argument.get("reader"),
                     include=argument.get("include"),
                     exclude=argument.get("exclude"),
-                    upsert=argument.get("upsert", False),
-                    skip_if_exists=argument.get("skip_if_exists", False),
+                    upsert=argument.get("upsert", upsert),
+                    skip_if_exists=argument.get("skip_if_exists", skip_if_exists),
                     remote_content=argument.get("remote_content", None),
                 )
 
@@ -102,7 +104,6 @@ class Knowledge:
             upsert = kwargs.get("upsert", False)
             skip_if_exists = kwargs.get("skip_if_exists", False)
             remote_content = kwargs.get("remote_content", None)
-
             for path in paths:
                 await self.add_content_async(
                     name=name,

agno/knowledge/reader/website_reader.py

@@ -106,23 +106,35 @@ class WebsiteReader(Reader):
             """
             Check if the tag matches any of the relevant tags or class names
             """
-            if tag.name in ["article", "main"]:
+            if not isinstance(tag, Tag):
+                return False
+
+            if tag.name in ["article", "main", "section"]:
+                return True
+
+            classes = tag.get("class", [])
+            content_classes = ["content", "main-content", "post-content", "entry-content", "article-body"]
+            if any(cls in content_classes for cls in classes):
                 return True
-            if any(cls in ["content", "main-content", "post-content"] for cls in tag.get("class", [])):  # type: ignore
+
+            # Check for common content IDs
+            tag_id = tag.get("id", "")
+            if tag_id in ["content", "main", "article"]:
                 return True
+
             return False
 
-        # Use a single call to 'find' with a custom function to match tags or classes
+        # Try to find main content element
         element = soup.find(match)
         if element:
+            # Remove common unwanted elements from the found content
+            for unwanted in element.find_all(["script", "style", "nav", "header", "footer"]):
+                unwanted.decompose()
             return element.get_text(strip=True, separator=" ")
 
-        # If we only have a div without specific content classes, return empty string
-        if soup.find("div") and not any(
-            soup.find(class_=class_name) for class_name in ["content", "main-content", "post-content"]
-        ):
-            return ""
-
+        # Fallback: get full page content
+        for unwanted in soup.find_all(["script", "style", "nav", "header", "footer"]):
+            unwanted.decompose()
         return soup.get_text(strip=True, separator=" ")
 
     def crawl(self, url: str, starting_depth: int = 1) -> Dict[str, str]:
@@ -164,7 +176,7 @@ class WebsiteReader(Reader):
             if (
                 current_url in self._visited
                 or not urlparse(current_url).netloc.endswith(primary_domain)
-                or current_depth > self.max_depth
+                or (current_depth > self.max_depth and current_url != url)
                 or num_links >= self.max_links
             ):
                 continue
@@ -174,13 +186,14 @@ class WebsiteReader(Reader):
 
             try:
                 log_debug(f"Crawling: {current_url}")
+
                 response = (
-                    httpx.get(current_url, timeout=self.timeout, proxy=self.proxy)
+                    httpx.get(current_url, timeout=self.timeout, proxy=self.proxy, follow_redirects=True)
                     if self.proxy
-                    else httpx.get(current_url, timeout=self.timeout)
+                    else httpx.get(current_url, timeout=self.timeout, follow_redirects=True)
                 )
-
                 response.raise_for_status()
+
                 soup = BeautifulSoup(response.content, "html.parser")
 
                 # Extract main content
@@ -213,9 +226,13 @@ class WebsiteReader(Reader):
 
             except httpx.HTTPStatusError as e:
                 # Log HTTP status errors but continue crawling other pages
-                logger.warning(f"HTTP status error while crawling {current_url}: {e}")
-                # For the initial URL, we should raise the error
-                if current_url == url and not crawler_result:
+                # Skip redirect errors (3xx) as they should be handled by follow_redirects
+                if e.response.status_code >= 300 and e.response.status_code < 400:
+                    logger.debug(f"Redirect encountered for {current_url}, skipping: {e}")
+                else:
+                    logger.warning(f"HTTP status error while crawling {current_url}: {e}")
+                # For the initial URL, we should raise the error only if it's not a redirect
+                if current_url == url and not crawler_result and not (300 <= e.response.status_code < 400):
                     raise
             except httpx.RequestError as e:
                 # Log request errors but continue crawling other pages