agents.react_class.react_agent2.dynamic_agent¶

Dynamic React Agent - an extension of React agent with tool selection capabilities.

Classes¶

`DynamicReactAgent`	A React agent with dynamic tool selection.
`DynamicReactAgentConfig`	Configuration for a React agent with dynamic tool selection.
`DynamicReactAgentState`	Extended schema for dynamic tool selection.

Functions¶

create_dynamic_react_agent(tools[, system_prompt, ...])

Create a dynamic react agent with minimal configuration.

Module Contents¶

class agents.react_class.react_agent2.dynamic_agent.DynamicReactAgent(config)¶

Bases: haive.agents.react_class.react_agent2.agent2.ReactAgent

A React agent with dynamic tool selection.

This agent extends the React pattern with dynamic tool selection, making it efficient when dealing with a large number of tools.

Initialize the dynamic react agent.

Parameters:: config (DynamicReactAgentConfig)

register_tools(tools)¶

Parameters:: tools (list[langchain_core.tools.BaseTool]) – List of tools to register
Return type:: None

run(input_data, **kwargs)¶

Run the agent with the given input.

Parameters:

input_data (str | list[str] | dict[str, Any] | pydantic.BaseModel) – Input data in various formats
**kwargs – Additional runtime configuration

Returns:

Final state or output

Return type:

dict[str, Any]

setup_workflow()¶

Set up the workflow with dynamic tool selection.

Return type:: None

property vector_store: Any¶

Get the vector store property.

Return type:: Any

class agents.react_class.react_agent2.dynamic_agent.DynamicReactAgentConfig¶

Bases: haive.agents.react_class.react_agent2.config2.ReactAgentConfig

Configuration for a React agent with dynamic tool selection.

This agent can handle a large number of tools by dynamically selecting which tools to make available to the LLM based on the context.

classmethod from_tools(tools, system_prompt=None, model='gpt-4o', temperature=0.7, name=None, max_tools_per_turn=5, max_iterations=10, response_format=None, use_memory=True, visualize=True, repeat_selection=True, **kwargs)¶

Create a DynamicReactAgentConfig from a list of tools.

Parameters:

tools (list[langchain_core.tools.BaseTool]) – List of tools to use
system_prompt (str | None) – Optional system prompt
model (str) – Model name
temperature (float) – Temperature for generation
name (str | None) – Optional agent name
max_tools_per_turn (int) – Maximum number of tools to select per turn
max_iterations (int) – Maximum iterations for React agent
response_format (type[pydantic.BaseModel] | dict[str, Any] | None) – Optional structured output model
use_memory (bool) – Whether to use memory
visualize (bool) – Whether to visualize the graph
repeat_selection (bool) – Whether to repeat tool selection after each tool invocation
**kwargs – Additional kwargs for the config

Returns:

DynamicReactAgentConfig instance

Return type:

DynamicReactAgentConfig

class agents.react_class.react_agent2.dynamic_agent.DynamicReactAgentState(/, **data)¶

Bases: haive.agents.react_class.react_agent2.state2.ReactAgentState

Extended schema for dynamic tool selection.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

model_config¶: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

agents.react_class.react_agent2.dynamic_agent.create_dynamic_react_agent(tools, system_prompt=None, model='gpt-4o', temperature=0.7, name=None, max_tools_per_turn=5, max_iterations=10, response_format=None, use_memory=True, visualize=True, repeat_selection=True, vector_store_config=None, tool_documents=None, **kwargs)¶

Create a dynamic react agent with minimal configuration.

Parameters:

tools (list[langchain_core.tools.BaseTool]) – List of tools the agent can use
system_prompt (str | None) – Optional system prompt
model (str) – Model name to use
temperature (float) – Temperature for generation
name (str | None) – Optional agent name
max_tools_per_turn (int) – Maximum number of tools to select per turn
max_iterations (int) – Maximum iterations for React agent
response_format (type[pydantic.BaseModel] | dict[str, Any] | None) – Optional structured output model
use_memory (bool) – Whether to use memory
visualize (bool) – Whether to visualize the graph
repeat_selection (bool) – Whether to repeat tool selection after each tool invocation
vector_store_config (haive.core.models.vectorstore.base.VectorStoreConfig | None) – Optional vector store configuration
tool_documents (list[langchain_core.documents.Document] | None) – Optional pre-created tool documents
**kwargs – Additional configuration parameters

Returns:

DynamicReactAgent instance

Return type:

DynamicReactAgent