Updates the content and memory for a given node based on the sender type. Args: node_name (`str`): The name of the n

Log error to log file Args: message: Error message

Parses and summarizes the search results with extracted references. If the search engine summary doesn't contain enough info, the c

Execute the tool with given parameters.

Store a failed URL attempt

Save failed URLs to file

Save URL mapping to file

Generates the request data for different model types. Args: model_type (str): The type of the model (e.g., 'gpt', 'i

Get URL content, reading from cache if available, otherwise fetching from web Args: url: URL to fetch

Generates the current plan and subqueries based on the user input and recorder. Args: message (`Union[List[U

Generate completion from a list of templates. Args: messages (List[dict]): a list of prompt dictionaries gen_para

Generates references from the search result graph and updates the reference indices. Args: query (`str`): The origin

Generate unique filename for URL Args: url: URL Returns: Filename string

Store URL content as a separate JSON file Args: url: URL to store content: Content to store

Generate completion from a list of templates. Args: messages (List[dict]): a list of prompt dictionaries gen_para

Updates references within a given string based on reference indices and the provided pointer. Args: ref (`str`): The

Add a search sub-question node. Args: node_name (str): Name of the node. node_content (str): Content of the sub-q

Add a response node. If the current information meets the query requirements, add a response node. Args: thought (str

Add the root node to the graph. Args: node_content (str): Content of the node. node_name (str, optional): Name of

Generate responses given the contexts. Args: inputs (Union[List[dict], List[List[dict]]]): a list of messages

Create an instance based on the configuration where 'type' is a preserved key to indicate the class (path). When accepting non-dictionary in

Generate results given a str (or list of) inputs. Args: inputs (Union[str, List[str]]): gen_params (dict): The in

Executes the search process iteratively, refining the search results over multiple turns. Args: query (`str`): The u

DOCX file parsing method

Excel file parsing method

Reset the graph by clearing all nodes and edges.

Manually store URL content to cache Args: url: URL to store content: Content to store

Generate responses given the contexts. Args: inputs (List[dict]): a list of messages gen_params: additional gener

Convert tool to function call format.

Execute the tool with given parameters.

Initialize the caching mechanism Args: cache_dir: Cache directory path url_map_file: URL mappin

Initializes the Recorder object to track query content and memory. Args: action (`str`): The action to be performed

A module responsible for parsing and summarizing relevant information from search results. Args: user_input_template

Initializes the SearchAgent with required components for each stage of the search process. Args: planner (`P

Add an edge between two nodes. Args: start_node (str): Name of the starting node. end_node (str): Name of the end

Generate completion from a list of templates. Args: inputs (Union[List[dict], List[List[dict]]]): gen_params (dic