agno 2.2.9__py3-none-any.whl → 2.2.11__py3-none-any.whl

agno/tools/parallel.py

@@ -0,0 +1,314 @@
+import json
+from os import getenv
+from typing import Any, Dict, List, Optional
+
+from agno.tools import Toolkit
+from agno.utils.log import log_error
+
+try:
+    from parallel import Parallel as ParallelClient
+except ImportError:
+    raise ImportError("`parallel-web` not installed. Please install using `pip install parallel-web`")
+
+
+class CustomJSONEncoder(json.JSONEncoder):
+    """Custom JSON encoder that handles non-serializable types by converting them to strings."""
+
+    def default(self, obj):
+        try:
+            return super().default(obj)
+        except TypeError:
+            return str(obj)
+
+
+class ParallelTools(Toolkit):
+    """
+    ParallelTools provides access to Parallel's web search and extraction APIs.
+
+    Parallel offers powerful APIs optimized for AI agents:
+    - Search API: AI-optimized web search that returns relevant excerpts tailored for LLMs
+    - Extract API: Extract content from specific URLs in clean markdown format, handling JavaScript-heavy pages and PDFs
+
+    Args:
+        api_key (Optional[str]): Parallel API key. If not provided, will use PARALLEL_API_KEY environment variable.
+        enable_search (bool): Enable Search API functionality. Default is True.
+        enable_extract (bool): Enable Extract API functionality. Default is True.
+        all (bool): Enable all tools. Overrides individual flags when True. Default is False.
+        max_results (int): Default maximum number of results for search operations. Default is 10.
+        max_chars_per_result (int): Default maximum characters per result for search operations. Default is 10000.
+        beta_version (str): Beta API version header. Default is "search-extract-2025-10-10".
+        mode (Optional[str]): Default search mode. Options: "one-shot" or "agentic". Default is None.
+        include_domains (Optional[List[str]]): Default domains to restrict results to. Default is None.
+        exclude_domains (Optional[List[str]]): Default domains to exclude from results. Default is None.
+        max_age_seconds (Optional[int]): Default cache age threshold (minimum 600). Default is None.
+        timeout_seconds (Optional[float]): Default timeout for content retrieval. Default is None.
+        disable_cache_fallback (Optional[bool]): Default cache fallback behavior. Default is None.
+    """
+
+    def __init__(
+        self,
+        api_key: Optional[str] = None,
+        enable_search: bool = True,
+        enable_extract: bool = True,
+        all: bool = False,
+        max_results: int = 10,
+        max_chars_per_result: int = 10000,
+        beta_version: str = "search-extract-2025-10-10",
+        mode: Optional[str] = None,
+        include_domains: Optional[List[str]] = None,
+        exclude_domains: Optional[List[str]] = None,
+        max_age_seconds: Optional[int] = None,
+        timeout_seconds: Optional[float] = None,
+        disable_cache_fallback: Optional[bool] = None,
+        **kwargs,
+    ):
+        self.api_key: Optional[str] = api_key or getenv("PARALLEL_API_KEY")
+        if not self.api_key:
+            log_error("PARALLEL_API_KEY not set. Please set the PARALLEL_API_KEY environment variable.")
+
+        self.max_results = max_results
+        self.max_chars_per_result = max_chars_per_result
+        self.beta_version = beta_version
+        self.mode = mode
+        self.include_domains = include_domains
+        self.exclude_domains = exclude_domains
+        self.max_age_seconds = max_age_seconds
+        self.timeout_seconds = timeout_seconds
+        self.disable_cache_fallback = disable_cache_fallback
+
+        self.parallel_client = ParallelClient(
+            api_key=self.api_key, default_headers={"parallel-beta": self.beta_version}
+        )
+
+        tools: List[Any] = []
+        if all or enable_search:
+            tools.append(self.parallel_search)
+        if all or enable_extract:
+            tools.append(self.parallel_extract)
+
+        super().__init__(name="parallel_tools", tools=tools, **kwargs)
+
+    def parallel_search(
+        self,
+        objective: Optional[str] = None,
+        search_queries: Optional[List[str]] = None,
+        max_results: Optional[int] = None,
+        max_chars_per_result: Optional[int] = None,
+    ) -> str:
+        """Use this function to search the web using Parallel's Search API with a natural language objective.
+        You must provide at least one of objective or search_queries.
+
+        Args:
+            objective (Optional[str]): Natural-language description of what the web search is trying to find.
+            search_queries (Optional[List[str]]): Traditional keyword queries with optional search operators.
+            max_results (Optional[int]): Upper bound on results returned. Overrides constructor default.
+            max_chars_per_result (Optional[int]): Upper bound on total characters per url for excerpts.
+
+        Returns:
+            str: A JSON formatted string containing the search results with URLs, titles, publish dates, and relevant excerpts.
+        """
+        try:
+            if not objective and not search_queries:
+                return json.dumps({"error": "Please provide at least one of: objective or search_queries"}, indent=2)
+
+            # Use instance defaults if not provided
+            final_max_results = max_results if max_results is not None else self.max_results
+
+            search_params: Dict[str, Any] = {
+                "max_results": final_max_results,
+            }
+
+            # Add objective if provided
+            if objective:
+                search_params["objective"] = objective
+
+            # Add search_queries if provided
+            if search_queries:
+                search_params["search_queries"] = search_queries
+
+            # Add mode from constructor default
+            if self.mode:
+                search_params["mode"] = self.mode
+
+            # Add excerpts configuration
+            excerpts_config: Dict[str, Any] = {}
+            final_max_chars = max_chars_per_result if max_chars_per_result is not None else self.max_chars_per_result
+            if final_max_chars is not None:
+                excerpts_config["max_chars_per_result"] = final_max_chars
+
+            if excerpts_config:
+                search_params["excerpts"] = excerpts_config
+
+            # Add source_policy from constructor defaults
+            source_policy: Dict[str, Any] = {}
+            if self.include_domains:
+                source_policy["include_domains"] = self.include_domains
+            if self.exclude_domains:
+                source_policy["exclude_domains"] = self.exclude_domains
+
+            if source_policy:
+                search_params["source_policy"] = source_policy
+
+            # Add fetch_policy from constructor defaults
+            fetch_policy: Dict[str, Any] = {}
+            if self.max_age_seconds is not None:
+                fetch_policy["max_age_seconds"] = self.max_age_seconds
+            if self.timeout_seconds is not None:
+                fetch_policy["timeout_seconds"] = self.timeout_seconds
+            if self.disable_cache_fallback is not None:
+                fetch_policy["disable_cache_fallback"] = self.disable_cache_fallback
+
+            if fetch_policy:
+                search_params["fetch_policy"] = fetch_policy
+
+            search_result = self.parallel_client.beta.search(**search_params)
+
+            # Use model_dump() if available, otherwise convert to dict
+            try:
+                if hasattr(search_result, "model_dump"):
+                    return json.dumps(search_result.model_dump(), cls=CustomJSONEncoder)
+            except Exception:
+                pass
+
+            # Manually format the results
+            formatted_results: Dict[str, Any] = {
+                "search_id": getattr(search_result, "search_id", ""),
+                "results": [],
+            }
+
+            if hasattr(search_result, "results") and search_result.results:
+                results_list: List[Dict[str, Any]] = []
+                for result in search_result.results:
+                    formatted_result: Dict[str, Any] = {
+                        "title": getattr(result, "title", ""),
+                        "url": getattr(result, "url", ""),
+                        "publish_date": getattr(result, "publish_date", ""),
+                        "excerpt": getattr(result, "excerpt", ""),
+                    }
+                    results_list.append(formatted_result)
+                formatted_results["results"] = results_list
+
+            if hasattr(search_result, "warnings"):
+                formatted_results["warnings"] = search_result.warnings
+
+            if hasattr(search_result, "usage"):
+                formatted_results["usage"] = search_result.usage
+
+            return json.dumps(formatted_results, cls=CustomJSONEncoder, indent=2)
+
+        except Exception as e:
+            log_error(f"Error searching Parallel for objective '{objective}': {e}")
+            return json.dumps({"error": f"Search failed: {str(e)}"}, indent=2)
+
+    def parallel_extract(
+        self,
+        urls: List[str],
+        objective: Optional[str] = None,
+        search_queries: Optional[List[str]] = None,
+        excerpts: bool = True,
+        max_chars_per_excerpt: Optional[int] = None,
+        full_content: bool = False,
+        max_chars_for_full_content: Optional[int] = None,
+    ) -> str:
+        """Use this function to extract content from specific URLs using Parallel's Extract API.
+
+        Args:
+            urls (List[str]): List of public URLs to extract content from.
+            objective (Optional[str]): Search focus to guide content extraction.
+            search_queries (Optional[List[str]]): Keywords for targeting relevant content.
+            excerpts (bool): Include relevant text snippets.
+            max_chars_per_excerpt (Optional[int]): Upper bound on total characters per url. Only used when excerpts is True.
+            full_content (bool): Include complete page text.
+            max_chars_for_full_content (Optional[int]): Limit on characters per url. Only used when full_content is True.
+
+        Returns:
+            str: A JSON formatted string containing extracted content with titles, publish dates, excerpts and/or full content.
+        """
+        try:
+            if not urls:
+                return json.dumps({"error": "Please provide at least one URL to extract"}, indent=2)
+
+            extract_params: Dict[str, Any] = {
+                "urls": urls,
+            }
+
+            # Add objective if provided
+            if objective:
+                extract_params["objective"] = objective
+
+            # Add search_queries if provided
+            if search_queries:
+                extract_params["search_queries"] = search_queries
+
+            # Add excerpts configuration
+            if excerpts and max_chars_per_excerpt is not None:
+                extract_params["excerpts"] = {"max_chars_per_result": max_chars_per_excerpt}
+            else:
+                extract_params["excerpts"] = excerpts
+
+            # Add full_content configuration
+            if full_content and max_chars_for_full_content is not None:
+                extract_params["full_content"] = {"max_chars_per_result": max_chars_for_full_content}
+            else:
+                extract_params["full_content"] = full_content
+
+            # Add fetch_policy from constructor defaults
+            fetch_policy: Dict[str, Any] = {}
+            if self.max_age_seconds is not None:
+                fetch_policy["max_age_seconds"] = self.max_age_seconds
+            if self.timeout_seconds is not None:
+                fetch_policy["timeout_seconds"] = self.timeout_seconds
+            if self.disable_cache_fallback is not None:
+                fetch_policy["disable_cache_fallback"] = self.disable_cache_fallback
+
+            if fetch_policy:
+                extract_params["fetch_policy"] = fetch_policy
+
+            extract_result = self.parallel_client.beta.extract(**extract_params)
+
+            # Use model_dump() if available, otherwise convert to dict
+            try:
+                if hasattr(extract_result, "model_dump"):
+                    return json.dumps(extract_result.model_dump(), cls=CustomJSONEncoder)
+            except Exception:
+                pass
+
+            # Manually format the results
+            formatted_results: Dict[str, Any] = {
+                "extract_id": getattr(extract_result, "extract_id", ""),
+                "results": [],
+                "errors": [],
+            }
+
+            if hasattr(extract_result, "results") and extract_result.results:
+                results_list: List[Dict[str, Any]] = []
+                for result in extract_result.results:
+                    formatted_result: Dict[str, Any] = {
+                        "url": getattr(result, "url", ""),
+                        "title": getattr(result, "title", ""),
+                        "publish_date": getattr(result, "publish_date", ""),
+                    }
+
+                    if excerpts and hasattr(result, "excerpts"):
+                        formatted_result["excerpts"] = result.excerpts
+
+                    if full_content and hasattr(result, "full_content"):
+                        formatted_result["full_content"] = result.full_content
+
+                    results_list.append(formatted_result)
+                formatted_results["results"] = results_list
+
+            if hasattr(extract_result, "errors") and extract_result.errors:
+                formatted_results["errors"] = extract_result.errors
+
+            if hasattr(extract_result, "warnings"):
+                formatted_results["warnings"] = extract_result.warnings
+
+            if hasattr(extract_result, "usage"):
+                formatted_results["usage"] = extract_result.usage
+
+            return json.dumps(formatted_results, cls=CustomJSONEncoder, indent=2)
+
+        except Exception as e:
+            log_error(f"Error extracting from Parallel: {e}")
+            return json.dumps({"error": f"Extract failed: {str(e)}"}, indent=2)

agno/utils/models/claude.py

@@ -68,6 +68,8 @@ def _format_image_for_message(image: Image) -> Optional[Dict[str, Any]]:
     }
 
     try:
+        img_type = None
+
         # Case 0: Image is an Anthropic uploaded file
         if image.content is not None and hasattr(image.content, "id"):
             content_bytes = image.content
@@ -80,7 +82,6 @@ def _format_image_for_message(image: Image) -> Optional[Dict[str, Any]]:
             import os
             from urllib.parse import urlparse
 
-            img_type = None
             if image.url:
                 parsed_url = urlparse(image.url)
                 _, ext = os.path.splitext(parsed_url.path)

agno/workflow/agent.py

@@ -268,7 +268,7 @@ Guidelines:
                     user_id=session_from_db.user_id,
                     execution_input=workflow_execution_input,
                     workflow_run_response=workflow_run_response,
-                    session_state=run_context.session_state,
+                    run_context=run_context,
                     stream_events=True,
                     websocket_handler=websocket_handler,
                 ):
@@ -286,7 +286,7 @@ Guidelines:
                     user_id=session_from_db.user_id,
                     execution_input=workflow_execution_input,
                     workflow_run_response=workflow_run_response,
-                    session_state=run_context.session_state,
+                    run_context=run_context,
                 )
 
                 if isinstance(result.content, str):

agno/workflow/condition.py

@@ -4,6 +4,7 @@ from typing import Any, AsyncIterator, Awaitable, Callable, Dict, Iterator, List
 from uuid import uuid4
 
 from agno.run.agent import RunOutputEvent
+from agno.run.base import RunContext
 from agno.run.team import TeamRunOutputEvent
 from agno.run.workflow import (
     ConditionExecutionCompletedEvent,
@@ -175,6 +176,7 @@ class Condition:
         user_id: Optional[str] = None,
         workflow_run_response: Optional[WorkflowRunOutput] = None,
         store_executor_outputs: bool = True,
+        run_context: Optional[RunContext] = None,
         session_state: Optional[Dict[str, Any]] = None,
         workflow_session: Optional[WorkflowSession] = None,
         add_workflow_history_to_steps: Optional[bool] = False,
@@ -188,7 +190,11 @@ class Condition:
         self._prepare_steps()
 
         # Evaluate the condition
-        condition_result = self._evaluate_condition(step_input, session_state)
+        if run_context is not None and run_context.session_state is not None:
+            condition_result = self._evaluate_condition(step_input, session_state=run_context.session_state)
+        else:
+            condition_result = self._evaluate_condition(step_input, session_state=session_state)
+
         log_debug(f"Condition {self.name} evaluated to: {condition_result}")
 
         if not condition_result:
@@ -214,6 +220,7 @@ class Condition:
                     user_id=user_id,
                     workflow_run_response=workflow_run_response,
                     store_executor_outputs=store_executor_outputs,
+                    run_context=run_context,
                     session_state=session_state,
                     workflow_session=workflow_session,
                     add_workflow_history_to_steps=add_workflow_history_to_steps,
@@ -284,6 +291,7 @@ class Condition:
         workflow_run_response: Optional[WorkflowRunOutput] = None,
         step_index: Optional[Union[int, tuple]] = None,
         store_executor_outputs: bool = True,
+        run_context: Optional[RunContext] = None,
         session_state: Optional[Dict[str, Any]] = None,
         parent_step_id: Optional[str] = None,
         workflow_session: Optional[WorkflowSession] = None,
@@ -298,7 +306,10 @@ class Condition:
         self._prepare_steps()
 
         # Evaluate the condition
-        condition_result = self._evaluate_condition(step_input, session_state)
+        if run_context is not None and run_context.session_state is not None:
+            condition_result = self._evaluate_condition(step_input, session_state=run_context.session_state)
+        else:
+            condition_result = self._evaluate_condition(step_input, session_state=session_state)
         log_debug(f"Condition {self.name} evaluated to: {condition_result}")
 
         # Considering both stream_events and stream_intermediate_steps (deprecated)
@@ -363,6 +374,7 @@ class Condition:
                     workflow_run_response=workflow_run_response,
                     step_index=child_step_index,
                     store_executor_outputs=store_executor_outputs,
+                    run_context=run_context,
                     session_state=session_state,
                     parent_step_id=conditional_step_id,
                     workflow_session=workflow_session,
@@ -447,6 +459,7 @@ class Condition:
         user_id: Optional[str] = None,
         workflow_run_response: Optional[WorkflowRunOutput] = None,
         store_executor_outputs: bool = True,
+        run_context: Optional[RunContext] = None,
         session_state: Optional[Dict[str, Any]] = None,
         workflow_session: Optional[WorkflowSession] = None,
         add_workflow_history_to_steps: Optional[bool] = False,
@@ -460,7 +473,10 @@ class Condition:
         self._prepare_steps()
 
         # Evaluate the condition
-        condition_result = await self._aevaluate_condition(step_input, session_state)
+        if run_context is not None and run_context.session_state is not None:
+            condition_result = await self._aevaluate_condition(step_input, session_state=run_context.session_state)
+        else:
+            condition_result = await self._aevaluate_condition(step_input, session_state=session_state)
         log_debug(f"Condition {self.name} evaluated to: {condition_result}")
 
         if not condition_result:
@@ -488,6 +504,7 @@ class Condition:
                     user_id=user_id,
                     workflow_run_response=workflow_run_response,
                     store_executor_outputs=store_executor_outputs,
+                    run_context=run_context,
                     session_state=session_state,
                     workflow_session=workflow_session,
                     add_workflow_history_to_steps=add_workflow_history_to_steps,
@@ -556,6 +573,7 @@ class Condition:
         workflow_run_response: Optional[WorkflowRunOutput] = None,
         step_index: Optional[Union[int, tuple]] = None,
         store_executor_outputs: bool = True,
+        run_context: Optional[RunContext] = None,
         session_state: Optional[Dict[str, Any]] = None,
         parent_step_id: Optional[str] = None,
         workflow_session: Optional[WorkflowSession] = None,
@@ -570,7 +588,10 @@ class Condition:
         self._prepare_steps()
 
         # Evaluate the condition
-        condition_result = await self._aevaluate_condition(step_input, session_state)
+        if run_context is not None and run_context.session_state is not None:
+            condition_result = await self._aevaluate_condition(step_input, session_state=run_context.session_state)
+        else:
+            condition_result = await self._aevaluate_condition(step_input, session_state=session_state)
         log_debug(f"Condition {self.name} evaluated to: {condition_result}")
 
         # Considering both stream_events and stream_intermediate_steps (deprecated)
@@ -637,6 +658,7 @@ class Condition:
                     workflow_run_response=workflow_run_response,
                     step_index=child_step_index,
                     store_executor_outputs=store_executor_outputs,
+                    run_context=run_context,
                     session_state=session_state,
                     parent_step_id=conditional_step_id,
                     workflow_session=workflow_session,

agno/workflow/loop.py

@@ -4,6 +4,7 @@ from typing import Any, AsyncIterator, Awaitable, Callable, Dict, Iterator, List
 from uuid import uuid4
 
 from agno.run.agent import RunOutputEvent
+from agno.run.base import RunContext
 from agno.run.team import TeamRunOutputEvent
 from agno.run.workflow import (
     LoopExecutionCompletedEvent,
@@ -132,6 +133,7 @@ class Loop:
         user_id: Optional[str] = None,
         workflow_run_response: Optional[WorkflowRunOutput] = None,
         store_executor_outputs: bool = True,
+        run_context: Optional[RunContext] = None,
         session_state: Optional[Dict[str, Any]] = None,
         workflow_session: Optional[WorkflowSession] = None,
         add_workflow_history_to_steps: Optional[bool] = False,
@@ -160,6 +162,7 @@ class Loop:
                     user_id=user_id,
                     workflow_run_response=workflow_run_response,
                     store_executor_outputs=store_executor_outputs,
+                    run_context=run_context,
                     session_state=session_state,
                     workflow_session=workflow_session,
                     add_workflow_history_to_steps=add_workflow_history_to_steps,
@@ -232,6 +235,7 @@ class Loop:
         workflow_run_response: Optional[WorkflowRunOutput] = None,
         step_index: Optional[Union[int, tuple]] = None,
         store_executor_outputs: bool = True,
+        run_context: Optional[RunContext] = None,
         session_state: Optional[Dict[str, Any]] = None,
         parent_step_id: Optional[str] = None,
         workflow_session: Optional[WorkflowSession] = None,
@@ -311,6 +315,7 @@ class Loop:
                     workflow_run_response=workflow_run_response,
                     step_index=composite_step_index,
                     store_executor_outputs=store_executor_outputs,
+                    run_context=run_context,
                     session_state=session_state,
                     parent_step_id=loop_step_id,
                     add_workflow_history_to_steps=add_workflow_history_to_steps,
@@ -428,6 +433,7 @@ class Loop:
         user_id: Optional[str] = None,
         workflow_run_response: Optional[WorkflowRunOutput] = None,
         store_executor_outputs: bool = True,
+        run_context: Optional[RunContext] = None,
         session_state: Optional[Dict[str, Any]] = None,
         workflow_session: Optional[WorkflowSession] = None,
         add_workflow_history_to_steps: Optional[bool] = False,
@@ -458,6 +464,7 @@ class Loop:
                     user_id=user_id,
                     workflow_run_response=workflow_run_response,
                     store_executor_outputs=store_executor_outputs,
+                    run_context=run_context,
                     session_state=session_state,
                     workflow_session=workflow_session,
                     add_workflow_history_to_steps=add_workflow_history_to_steps,
@@ -533,6 +540,7 @@ class Loop:
         workflow_run_response: Optional[WorkflowRunOutput] = None,
         step_index: Optional[Union[int, tuple]] = None,
         store_executor_outputs: bool = True,
+        run_context: Optional[RunContext] = None,
         session_state: Optional[Dict[str, Any]] = None,
         parent_step_id: Optional[str] = None,
         workflow_session: Optional[WorkflowSession] = None,
@@ -612,6 +620,7 @@ class Loop:
                     workflow_run_response=workflow_run_response,
                     step_index=composite_step_index,
                     store_executor_outputs=store_executor_outputs,
+                    run_context=run_context,
                     session_state=session_state,
                     parent_step_id=loop_step_id,
                     workflow_session=workflow_session,