🐛Fix: ensure runnability of samples/agents

samples/agents/bedrock_agent_langgraph.ipynb

@@ -137,21 +137,23 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import boto3\n",
     "import json\n",
     "import time\n",
     "import uuid\n",
     "\n",
+    "import boto3\n",
+    "\n",
+    "\n",
     "def _create_agent_role(\n",
-    "        agent_region,\n",
-    "        foundation_model\n",
+    "    agent_region,\n",
+    "    foundation_model,\n",
     ") -> str:\n",
     "    \"\"\"\n",
     "    Create agent resource role prior to creation of agent, at this point we do not have agentId, keep it as wildcard\n",
     "\n",
     "    Args:\n",
     "        agent_region: AWS region in which is the Agent if available\n",
-    "        foundational_model: The model used for inference in AWS BedrockAgents\n",
+    "        foundation_model: The model used for inference in AWS BedrockAgents\n",
     "    Returns:\n",
     "       Agent execution role arn\n",
     "    \"\"\"\n",
@@ -188,7 +190,9 @@
     "            ]\n",
     "        }\n",
     "        role_name = f'bedrock_agent_{uuid.uuid4()}'\n",
+    "\n",
     "        iam_client = boto3.client('iam')\n",
+    "\n",
     "        response = iam_client.create_role(\n",
     "            RoleName=role_name,\n",
     "            AssumeRolePolicyDocument=assume_role_policy_document,\n",
@@ -205,9 +209,13 @@
     "    except Exception as exception:\n",
     "        raise exception\n",
     "\n",
+    "session = boto3.session.Session()\n",
+    "region_name = session.region_name\n",
+    "\n",
     "agent_resource_role_arn = _create_agent_role(\n",
-    "            agent_region='us-west-2',\n",
-    "            foundation_model=foundation_model)\n",
+    "    agent_region=region_name,\n",
+    "    foundation_model=foundation_model,\n",
+    ")\n",
     "\n",
     "agent_resource_role_arn"
    ]
@@ -229,14 +237,15 @@
    "source": [
     "from langchain_aws.agents import BedrockAgentsRunnable\n",
     "\n",
+    "\n",
     "agent = BedrockAgentsRunnable.create_agent(\n",
-    "            agent_name=\"langgraph_interest_rate_agent\",\n",
-    "            agent_resource_role_arn=agent_resource_role_arn,\n",
-    "            foundation_model=foundational_model,\n",
-    "            instruction=\"\"\"\n",
-    "            You are an agent who helps with getting the mortgage rate based on the current asset valuation\"\"\",\n",
-    "            tools=tools,\n",
-    "        )\n",
+    "    agent_name=\"langgraph_interest_rate_agent\",\n",
+    "    agent_resource_role_arn=agent_resource_role_arn,\n",
+    "    foundation_model=foundation_model,\n",
+    "    instruction=\"\"\"\n",
+    "    You are an agent who helps with getting the mortgage rate based on the current asset valuation\"\"\",\n",
+    "    tools=tools,\n",
+    ")\n",
     "agent"
    ]
   },
@@ -256,10 +265,12 @@
    "outputs": [],
    "source": [
     "import operator\n",
-    "from typing import TypedDict, Annotated, Tuple\n",
+    "from typing import TypedDict, Annotated\n",
     "from typing import Union\n",
+    "\n",
     "from langchain_aws.agents import BedrockAgentAction, BedrockAgentFinish\n",
     "\n",
+    "\n",
     "class AgentState(TypedDict):\n",
     "    \"\"\"\n",
     "        Defines the state for the Graph\n",
@@ -287,20 +298,21 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "from langgraph.prebuilt.tool_executor import ToolExecutor\n",
+    "from langgraph.prebuilt import ToolNode\n",
+    "\n",
     "\n",
     "# the agent node\n",
     "def run_agent(data):\n",
     "    agent_outcome = agent.invoke(data)\n",
     "    return {\"output\": agent_outcome}\n",
     "\n",
     "# the tools node\n",
-    "tool_executor = ToolExecutor(tools)\n",
+    "tool_node = ToolNode(tools)\n",
     "\n",
     "def execute_tools(data):\n",
     "    # Get the most recent output - this is the key added in the `agent` above\n",
     "    agent_action = data[\"output\"]\n",
-    "    output = tool_executor.invoke(agent_action[0])\n",
+    "    output = tool_node.invoke(agent_action[0])\n",
     "    tuple_output = agent_action[0], output\n",
     "    print(f\"Tuple output is {tuple_output}\")\n",
     "    return {\"intermediate_steps\": [tuple_output]}"