From 20a4aee5c1a15cd0415c916bbcc7512a642a9a37 Mon Sep 17 00:00:00 2001
From: Eugene Yurtsev <eyurtsev@gmail.com>
Date: Wed, 10 Apr 2024 09:27:32 -0400
Subject: [PATCH] Add factory for regular tool using agents (#169)

add factory for regular tool using agents

---------

Co-authored-by: Chester Curme <chester.curme@gmail.com>
---
 .../tool_usage/agents/__init__.py             |  2 +
 .../tool_usage/agents/tool_using_agent.py     | 72 +++++++++++++++++++
 2 files changed, 74 insertions(+)
 create mode 100644 langchain_benchmarks/tool_usage/agents/tool_using_agent.py

diff --git a/langchain_benchmarks/tool_usage/agents/__init__.py b/langchain_benchmarks/tool_usage/agents/__init__.py
index dad7deb..c59133f 100644
--- a/langchain_benchmarks/tool_usage/agents/__init__.py
+++ b/langchain_benchmarks/tool_usage/agents/__init__.py
@@ -12,6 +12,7 @@
 from langchain_benchmarks.tool_usage.agents.runnable_agent import (
     CustomRunnableAgentFactory,
 )
+from langchain_benchmarks.tool_usage.agents.tool_using_agent import StandardAgentFactory
 
 __all__ = [
     "OpenAIAgentFactory",
@@ -20,4 +21,5 @@
     "CustomAgentFactory",
     "AnthropicToolUserFactory",
     "CustomRunnableAgentFactory",
+    "StandardAgentFactory",
 ]
diff --git a/langchain_benchmarks/tool_usage/agents/tool_using_agent.py b/langchain_benchmarks/tool_usage/agents/tool_using_agent.py
new file mode 100644
index 0000000..cc1de36
--- /dev/null
+++ b/langchain_benchmarks/tool_usage/agents/tool_using_agent.py
@@ -0,0 +1,72 @@
+"""Factory for creating agents.
+
+This is useful for agents that follow the standard LangChain tool format.
+"""
+from langchain.agents import AgentExecutor
+from langchain_core.language_models import BaseChatModel
+from langchain_core.prompts import ChatPromptTemplate
+from langchain_core.runnables import Runnable
+
+from langchain_benchmarks.schema import ToolUsageTask
+from langchain_benchmarks.tool_usage.agents.adapters import apply_agent_executor_adapter
+
+
+class StandardAgentFactory:
+    """A standard agent factory.
+
+    Use this factory with chat models that support the standard LangChain tool
+    calling API where the chat model populates the tool_calls attribute on AIMessage.
+    """
+
+    def __init__(
+        self,
+        task: ToolUsageTask,
+        model: BaseChatModel,
+        prompt: ChatPromptTemplate,
+    ) -> None:
+        """Create an agent factory for the given tool usage task.
+
+        Args:
+            task: The task to create an agent factory for
+            model: chat model to use, must support tool usage
+            prompt: This is a chat prompt at the moment.
+                Must include an agent_scratchpad
+
+                For example,
+
+                ChatPromptTemplate.from_messages(
+                    [
+                        ("system", "{instructions}"),
+                        ("human", "{input}"),
+                        MessagesPlaceholder("agent_scratchpad"),
+                    ]
+                )
+        """
+        self.task = task
+        self.model = model
+        self.prompt = prompt
+
+    def __call__(self) -> Runnable:
+        """Call the factory to create Runnable agent."""
+        # Temporarily import here until new langchain is released with create_tools_agent
+        from langchain.agents import create_tool_calling_agent
+
+        env = self.task.create_environment()
+
+        if "instructions" in self.prompt.input_variables:
+            finalized_prompt = self.prompt.partial(instructions=self.task.instructions)
+        else:
+            finalized_prompt = self.prompt
+
+        agent = create_tool_calling_agent(self.model, env.tools, finalized_prompt)
+
+        executor = AgentExecutor(
+            agent=agent,
+            tools=env.tools,
+            handle_parsing_errors=True,
+            return_intermediate_steps=True,
+        )
+
+        return apply_agent_executor_adapter(
+            executor, state_reader=env.read_state
+        ).with_config({"run_name": "Agent", "metadata": {"task": self.task.name}})