add tool calling section to doc notebook

libs/ai-endpoints/docs/chat/nvidia_ai_endpoints.ipynb

@@ -548,6 +548,75 @@
    "source": [
     "conversation.invoke(\"Tell me about yourself.\")[\"response\"]"
    ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f3cbbba0",
+   "metadata": {},
+   "source": [
+    "## Tool calling\n",
+    "\n",
+    "Starting in v0.2, `ChatNVIDIA` supports [bind_tools](https://api.python.langchain.com/en/latest/language_models/langchain_core.language_models.chat_models.BaseChatModel.html#langchain_core.language_models.chat_models.BaseChatModel.bind_tools).\n",
+    "\n",
+    "`ChatNVIDIA` provides integration with the variety of models on [build.nvidia.com](https://build.nvidia.com) as well as local NIMs. Not all these models are trained for tool calling. Be sure to select a model that does have tool calling for your experimention and applications."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6f7b535e",
+   "metadata": {},
+   "source": [
+    "You can get a list of models that are known to support tool calling with,"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "e36c8911",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "tool_models = [model for model in ChatNVIDIA.get_available_models() if model.supports_tools]\n",
+    "tool_models"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b01d75a7",
+   "metadata": {},
+   "source": [
+    "With a tool capable model,"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "bd54f174",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain_core.pydantic_v1 import Field\n",
+    "from langchain_core.tools import tool\n",
+    "\n",
+    "@tool\n",
+    "def get_current_weather(\n",
+    "    location: str = Field(..., description=\"The location to get the weather for.\")\n",
+    "):\n",
+    "    \"\"\"Get the current weather for a location.\"\"\"\n",
+    "    ...\n",
+    "\n",
+    "llm = ChatNVIDIA(model=tool_models[0].id).bind_tools(tools=[get_current_weather])\n",
+    "response = llm.invoke(\"What is the weather in Boston?\")\n",
+    "response.tool_calls"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e08df68c",
+   "metadata": {},
+   "source": [
+    "See [How to use chat models to call tools](https://python.langchain.com/v0.2/docs/how_to/tool_calling/) for additional examples."
+   ]
   }
  ],
  "metadata": {