community: add PanelCallbackHandler

docs/docs/integrations/callbacks/panel.ipynb

@@ -0,0 +1,259 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Panel\n",
+    "\n",
+    "> [Panel](https://panel.holoviz.org/index.html) is an open-source Python library designed to streamline the development of robust tools, dashboards, and complex applications entirely within Python. With a comprehensive philosophy,\n",
+    "> Panel integrates seamlessly with the PyData ecosystem, offering powerful, interactive data tables, visualizations, and much more, to unlock, visualize, share, and collaborate on your data for efficient workflows.\n",
+    "\n",
+    "In this guide, we will go over how to setup the `PanelCallbackHandler`. The `PanelCallbackHandler` is useful for rendering and streaming the chain of thought from Langchain objects like Tools, Agents, and Chains. It inherits from Langchain’s BaseCallbackHandler.\n",
+    "\n",
+    "Check out the panel-chat-examples docs to see more examples on how to use PanelCallbackHandler. If you have an example to demo, we’d love to add it to the panel-chat-examples gallery!\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Installation and Setup\n",
+    "\n",
+    "```bash\n",
+    "pip install langchain panel\n",
+    "```\n",
+    "\n",
+    "See full instructions in Panel's [Getting started documentation](https://panel.holoviz.org/getting_started/index.html).\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Basic chat with an LLM\n",
+    "\n",
+    "To get started:\n",
+    "\n",
+    "1. Define a chat callback, like `respond` here.\n",
+    "2. Pass the instance of a `ChatFeed` or `ChatInterface` to `PanelCallbackHandler`.\n",
+    "3. Pass the `callback_handler` as a list into `callbacks` when constructing or using Langchain objects like `ChatOpenAI` here."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import panel as pn\n",
+    "from langchain_community.callbacks import PanelCallbackHandler\n",
+    "from langchain_openai import ChatOpenAI\n",
+    "\n",
+    "pn.extension()\n",
+    "\n",
+    "\n",
+    "def respond(contents):\n",
+    "    llm.invoke(contents)\n",
+    "\n",
+    "\n",
+    "chat_interface = pn.chat.ChatInterface(callback=respond)\n",
+    "callback = PanelCallbackHandler(chat_interface)\n",
+    "llm = ChatOpenAI(model_name=\"gpt-4o-mini\", streaming=True, callbacks=[callback])\n",
+    "chat_interface"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "This example shows the response from the LLM only. A LLM by it self does not show any chain of thought. Later we will build an agent that uses tools. This will show chain of thought."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Async chat with an LLM\n",
+    "\n",
+    "Using `async` prevents blocking the main thread, enabling concurrent interactions with the app. This improves responsiveness and user experience.\n",
+    "\n",
+    "To do so:\n",
+    "\n",
+    "1. Prefix the function with `async`\n",
+    "2. Prefix the call with `await`\n",
+    "3. Use `ainvoke` instead of `invoke`"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import panel as pn\n",
+    "from langchain_community.callbacks import PanelCallbackHandler\n",
+    "from langchain_openai import ChatOpenAI\n",
+    "\n",
+    "pn.extension()\n",
+    "\n",
+    "\n",
+    "async def respond(contents):\n",
+    "    await llm.ainvoke(contents)\n",
+    "\n",
+    "\n",
+    "chat_interface = pn.chat.ChatInterface(callback=respond)\n",
+    "callback = PanelCallbackHandler(chat_interface)\n",
+    "llm = ChatOpenAI(model_name=\"gpt-4o-mini\", streaming=True, callbacks=[callback])\n",
+    "chat_interface"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Agents with Tools\n",
+    "\n",
+    "Agents and tools can also be used. Simply pass callback to the `AgentExecutor` and its `invoke` method."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import panel as pn\n",
+    "from langchain import hub\n",
+    "from langchain.agents import AgentExecutor, create_react_agent, load_tools\n",
+    "from langchain_community.callbacks import PanelCallbackHandler\n",
+    "from langchain_openai import ChatOpenAI\n",
+    "\n",
+    "pn.extension()\n",
+    "\n",
+    "\n",
+    "def respond(contents):\n",
+    "    agent_executor.invoke({\"input\": contents}, {\"callbacks\": [callback]})\n",
+    "\n",
+    "\n",
+    "chat_interface = pn.chat.ChatInterface(callback=respond)\n",
+    "callback = PanelCallbackHandler(chat_interface)\n",
+    "llm = ChatOpenAI(model_name=\"gpt-4o-mini\", streaming=True, callbacks=[callback])\n",
+    "tools = load_tools([\"ddg-search\"])\n",
+    "prompt = hub.pull(\"hwchase17/react\")\n",
+    "agent = create_react_agent(llm, tools, prompt)\n",
+    "agent_executor = AgentExecutor(agent=agent, tools=tools, callbacks=[callback])\n",
+    "\n",
+    "chat_interface"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Chain with Retrievers\n",
+    "\n",
+    "RAG is also possible; simply pass `callback` again. Then ask it what the secret number is!"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from uuid import uuid4\n",
+    "\n",
+    "import panel as pn\n",
+    "from langchain.prompts import ChatPromptTemplate\n",
+    "from langchain.schema.runnable import RunnablePassthrough\n",
+    "from langchain.text_splitter import CharacterTextSplitter\n",
+    "from langchain.vectorstores import Chroma\n",
+    "from langchain_community.callbacks import PanelCallbackHandler\n",
+    "from langchain_openai import ChatOpenAI, OpenAIEmbeddings\n",
+    "\n",
+    "TEXT = \"The secret number is 888.\"\n",
+    "\n",
+    "TEMPLATE = \"\"\"Answer the question based only on the following context:\n",
+    "\n",
+    "{context}\n",
+    "\n",
+    "Question: {question}\n",
+    "\"\"\"\n",
+    "\n",
+    "pn.extension(design=\"material\")\n",
+    "\n",
+    "\n",
+    "@pn.cache\n",
+    "def get_vector_store():\n",
+    "    text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=100)\n",
+    "    texts = text_splitter.split_text(TEXT)\n",
+    "    embeddings = OpenAIEmbeddings()\n",
+    "    db = Chroma.from_texts(texts, embeddings)\n",
+    "    return db\n",
+    "\n",
+    "\n",
+    "def get_chain(callbacks):\n",
+    "    retriever = db.as_retriever(callbacks=callbacks)\n",
+    "    model = ChatOpenAI(callbacks=callbacks, streaming=True)\n",
+    "\n",
+    "    def format_docs(docs):\n",
+    "        text = \"\\n\\n\".join([d.page_content for d in docs])\n",
+    "        return text\n",
+    "\n",
+    "    def hack(docs):\n",
+    "        # https://github.com/langchain-ai/langchain/issues/7290\n",
+    "        for callback in callbacks:\n",
+    "            callback.on_retriever_end(docs, run_id=uuid4())\n",
+    "        return docs\n",
+    "\n",
+    "    return (\n",
+    "        {\"context\": retriever | hack | format_docs, \"question\": RunnablePassthrough()}\n",
+    "        | prompt\n",
+    "        | model\n",
+    "    )\n",
+    "\n",
+    "\n",
+    "async def respond(contents):\n",
+    "    chain = get_chain(callbacks=[callback])\n",
+    "    await chain.ainvoke(contents)\n",
+    "\n",
+    "\n",
+    "db = get_vector_store()\n",
+    "prompt = ChatPromptTemplate.from_template(TEMPLATE)\n",
+    "chat_interface = pn.chat.ChatInterface(callback=respond)\n",
+    "callback = PanelCallbackHandler(chat_interface)\n",
+    "\n",
+    "chat_interface"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.8"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

libs/community/langchain_community/callbacks/__init__.py

@@ -60,6 +60,9 @@
     from langchain_community.callbacks.openai_info import (
         OpenAICallbackHandler,
     )
+    from langchain_community.callbacks.panel_callback import (
+        PanelCallbackHandler,
+    )
     from langchain_community.callbacks.promptlayer_callback import (
         PromptLayerCallbackHandler,
     )
@@ -105,6 +108,7 @@
     "LabelStudioCallbackHandler": "langchain_community.callbacks.labelstudio_callback",
     "MlflowCallbackHandler": "langchain_community.callbacks.mlflow_callback",
     "OpenAICallbackHandler": "langchain_community.callbacks.openai_info",
+    "PanelCallbackHandler": "langchain_community.callbacks.panel_callback",
     "PromptLayerCallbackHandler": "langchain_community.callbacks.promptlayer_callback",
     "SageMakerCallbackHandler": "langchain_community.callbacks.sagemaker_callback",
     "StreamlitCallbackHandler": "langchain_community.callbacks.streamlit",
@@ -143,6 +147,7 @@ def __getattr__(name: str) -> Any:
     "LabelStudioCallbackHandler",
     "MlflowCallbackHandler",
     "OpenAICallbackHandler",
+    "PanelCallbackHandler",
     "PromptLayerCallbackHandler",
     "SageMakerCallbackHandler",
     "StreamlitCallbackHandler",