diff --git a/learn/generation/langchain/handbook/03a-token-counter.ipynb b/learn/generation/langchain/handbook/03a-token-counter.ipynb index ef3ebe7d..e3f7c7f9 100644 --- a/learn/generation/langchain/handbook/03a-token-counter.ipynb +++ b/learn/generation/langchain/handbook/03a-token-counter.ipynb @@ -1,27 +1,27 @@ { "cells": [ { - "attachments": {}, "cell_type": "markdown", - "metadata": {}, + "metadata": { + "id": "gPuN7cmcL9cv" + }, "source": [ - "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/pinecone-io/examples/blob/master/learn/generation/langchain/handbook/03a-token-counter.ipynb) [![Open nbviewer](https://raw.githubusercontent.com/pinecone-io/examples/master/assets/nbviewer-shield.svg)](https://nbviewer.org/github/pinecone-io/examples/blob/master/learn/generation/langchain/handbook/03a-token-counter.ipynb)" + "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/pinecone-io/examples/blob/master/learn/generation/langchain/handbook/03-langchain-conversational-memory.ipynb) [![Open nbviewer](https://raw.githubusercontent.com/pinecone-io/examples/master/assets/nbviewer-shield.svg)](https://nbviewer.org/github/pinecone-io/examples/blob/master/learn/generation/langchain/handbook/03-langchain-conversational-memory.ipynb)" ] }, { - "attachments": {}, "cell_type": "markdown", "metadata": { "id": "ytU8AV9uMty3" }, "source": [ - "#### [LangChain Handbook](https://pinecone.io/learn/langchain)\n", + "#### [LangChain Handbook](https://www.pinecone.io/learn/series/langchain/)\n", "\n", "# Conversational Memory\n", "\n", "## Extra Material: Token Counter\n", "\n", - "This is an additional piece of material alongside the [LangChain Handbook notebook on Conversational Memory](https://github.com/pinecone-io/examples/blob/master/generation/langchain/handbook/03-langchain-conversational-memory.ipynb).\n", + "This is an additional piece of material alongside the [LangChain Handbook notebook on Conversational Memory](https://github.com/pinecone-io/examples/blob/master/learn/generation/langchain/handbook/03-langchain-conversational-memory.ipynb).\n", "\n", "In this notebook we will count the number of tokens used in a conversation for different conversational memory types.\n", "\n", @@ -35,108 +35,89 @@ "colab": { "base_uri": "https://localhost:8080/" }, - "id": "9iK0npZ4Mgae", - "outputId": "f79e4b83-b1f2-4610-b032-6fbb042c3872" + "id": "zwS2JId1L9cv", + "outputId": "9a7c88d4-3e2f-4187-8b77-acda30622174" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ - "\u001b[2K \u001b[90m\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u001b[0m \u001b[32m6.3/6.3 MB\u001b[0m \u001b[31m23.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", - "\u001b[2K \u001b[90m\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u001b[0m \u001b[32m7.6/7.6 MB\u001b[0m \u001b[31m48.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", - "\u001b[2K \u001b[90m\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u2501\u001b[0m \u001b[32m190.3/190.3 KB\u001b[0m \u001b[31m6.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25l \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m0.0/1.0 MB\u001b[0m \u001b[31m?\u001b[0m eta \u001b[36m-:--:--\u001b[0m\r\u001b[2K \u001b[91m━━━━━━━━━━━━━\u001b[0m\u001b[90m╺\u001b[0m\u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m0.3/1.0 MB\u001b[0m \u001b[31m11.6 MB/s\u001b[0m eta \u001b[36m0:00:01\u001b[0m\r\u001b[2K \u001b[91m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[91m╸\u001b[0m \u001b[32m1.0/1.0 MB\u001b[0m \u001b[31m18.6 MB/s\u001b[0m eta \u001b[36m0:00:01\u001b[0m\r\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.0/1.0 MB\u001b[0m \u001b[31m14.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25h\u001b[?25l \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m0.0/2.5 MB\u001b[0m \u001b[31m?\u001b[0m eta \u001b[36m-:--:--\u001b[0m\r\u001b[2K \u001b[91m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[91m╸\u001b[0m \u001b[32m2.5/2.5 MB\u001b[0m \u001b[31m91.5 MB/s\u001b[0m eta \u001b[36m0:00:01\u001b[0m\r\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.5/2.5 MB\u001b[0m \u001b[31m47.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m65.3/65.3 kB\u001b[0m \u001b[31m2.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m363.0/363.0 kB\u001b[0m \u001b[31m20.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m45.2/45.2 kB\u001b[0m \u001b[31m2.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m50.9/50.9 kB\u001b[0m \u001b[31m1.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", "\u001b[?25h" ] } ], "source": [ - "!pip install -qU langchain openai transformers" + "!pip install -qU \\\n", + " langchain==0.3.25 \\\n", + " langchain-community==0.3.25 \\\n", + " langchain-openai==0.3.22 \\\n", + " transformers==4.52.4 \\\n", + " seaborn==0.13.2" ] }, { - "attachments": {}, - "cell_type": "markdown", - "metadata": { - "id": "g6-xDqLGRYsh" - }, - "source": [ - "Import required libraries and objects:" - ] - }, - { - "cell_type": "code", - "execution_count": 2, - "metadata": { - "id": "4nzzNSCxRcHS" - }, - "outputs": [], - "source": [ - "from getpass import getpass\n", - "\n", - "import openai\n", - "from langchain import OpenAI\n", - "from langchain.chains import LLMChain, ConversationChain\n", - "from langchain.chains.conversation.memory import (\n", - " ConversationBufferMemory,\n", - " ConversationSummaryMemory,\n", - " ConversationBufferWindowMemory,\n", - " ConversationSummaryBufferMemory\n", - ")\n", - "from langchain.callbacks import get_openai_callback\n", - "from tqdm.auto import tqdm" - ] - }, - { - "attachments": {}, "cell_type": "markdown", "metadata": { "id": "2zWvthQzUUzC" }, "source": [ - "To run the notebook we'll use OpenAI's `gpt-3.5-turbo` model. We initialize it via LangChain like so:" + "To run the notebook we'll use OpenAI's `gpt-4.1-mini` model. We initialize it via LangChain like so:" ] }, { "cell_type": "code", - "execution_count": 3, + "execution_count": 2, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "cvimTSA4Ublb", - "outputId": "075282fa-b447-4e87-9d31-bf5d8966fc08" + "outputId": "d600eb02-603c-448a-e223-8cb17974e8f5" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ - "\u00b7\u00b7\u00b7\u00b7\u00b7\u00b7\u00b7\u00b7\u00b7\u00b7\n" + "Enter your OpenAI API key: ··········\n" ] } ], "source": [ - "OPENAI_API_KEY = getpass()" + "import os\n", + "from getpass import getpass\n", + "\n", + "os.environ[\"OPENAI_API_KEY\"] = os.getenv(\"OPENAI_API_KEY\") \\\n", + " or getpass(\"Enter your OpenAI API key: \")\n", + "\n", + "OPENAI_API_KEY = os.getenv(\"OPENAI_API_KEY\")" ] }, { "cell_type": "code", - "execution_count": 4, + "execution_count": 3, "metadata": { "id": "oRjdCaSaUdCd" }, "outputs": [], "source": [ - "llm = OpenAI(\n", - " temperature=0, \n", + "from langchain_openai import ChatOpenAI\n", + "\n", + "llm = ChatOpenAI(\n", + " temperature=0,\n", " openai_api_key=OPENAI_API_KEY,\n", - " model_name='gpt-3.5-turbo' # can be used with llms like 'text-davinci-003'\n", + " model_name='gpt-4.1-mini'\n", ")" ] }, { - "attachments": {}, "cell_type": "markdown", "metadata": { "id": "fUv5H5SWUgNt" @@ -147,15 +128,27 @@ }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 4, "metadata": { "id": "EqTuGsENUnAr" }, "outputs": [], "source": [ - "def count_tokens(chain, query):\n", + "from langchain.callbacks import get_openai_callback\n", + "\n", + "def count_tokens(chain, query, config=None):\n", " with get_openai_callback() as cb:\n", - " result = chain.run(query)\n", + " # Handle both dict and string inputs\n", + " if isinstance(query, str):\n", + " query = {\"query\": query}\n", + "\n", + " # Use provided config or default\n", + " if config is None:\n", + " config = {\"configurable\": {\"session_id\": \"default\"}}\n", + "\n", + " result = chain.invoke(query, config=config)\n", + " print(f'Spent a total of {cb.total_tokens} tokens')\n", + "\n", " return {\n", " 'result': result,\n", " 'token_count': cb.total_tokens\n", @@ -163,23 +156,496 @@ ] }, { - "attachments": {}, "cell_type": "markdown", "metadata": { - "id": "OvCBfg9gUst-" + "id": "wH3HtA2oL9cw" }, "source": [ - "Let's define the conversation function:" + "## Define System Prompt and LCEL Pipeline" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "metadata": { + "id": "terrcggzL9cw" + }, + "outputs": [], + "source": [ + "from langchain.prompts import (\n", + " ChatPromptTemplate,\n", + " SystemMessagePromptTemplate,\n", + " HumanMessagePromptTemplate,\n", + " MessagesPlaceholder\n", + ")\n", + "from langchain.schema.output_parser import StrOutputParser\n", + "\n", + "# Define the prompt template\n", + "system_prompt = \"\"\"The following is a friendly conversation between a human and an AI. The AI is talkative and provides lots of specific details from its context. If the AI does not know the answer to a question, it truthfully says it does not know.\"\"\"\n", + "\n", + "prompt_template = ChatPromptTemplate.from_messages([\n", + " SystemMessagePromptTemplate.from_template(system_prompt),\n", + " MessagesPlaceholder(variable_name=\"history\"),\n", + " HumanMessagePromptTemplate.from_template(\"{query}\"),\n", + "])\n", + "\n", + "# Create the LCEL pipeline\n", + "output_parser = StrOutputParser()\n", + "pipeline = prompt_template | llm | output_parser" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "eQHFpyi3L9cw" + }, + "source": [ + "## Runnables with Message Histories" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "blRPHXkJL9cx" + }, + "source": [ + "Here the Runnable Classes which utilize different types of memory are defined.\n", + "\n", + "### Memory Type #1: Buffer Memory - Store the Entire Chat History\n", + "\n", + "An alternative to `ConversationBufferMemory`. The simplest method, which stores the entire chat history as memory." ] }, { "cell_type": "code", "execution_count": 6, "metadata": { - "id": "tEjQ7YZwU_rM" + "id": "iGPU-bOaL9cx" + }, + "outputs": [], + "source": [ + "from langchain_core.chat_history import InMemoryChatMessageHistory\n", + "\n", + "# Create a simple chat history storage\n", + "chat_map = {}\n", + "\n", + "def get_chat_history(session_id: str) -> InMemoryChatMessageHistory:\n", + " if session_id not in chat_map:\n", + " # if session ID doesn't exist, create a new chat history\n", + " chat_map[session_id] = InMemoryChatMessageHistory()\n", + " return chat_map[session_id]" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "m21c0CNYL9cx" + }, + "source": [ + "### Memory type #2: Summary - Store Summaries of Past Interactions\n", + "\n", + "This is an LCEL-Comptaible alternative to `ConversationSummaryMemory`. We keep a summary of our previous conversation snippets as our history. The summarization is performed by an LLM." + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "metadata": { + "id": "CdFI6MEXL9cx" }, "outputs": [], "source": [ + "from langchain_core.messages import BaseMessage, SystemMessage\n", + "\n", + "from langchain_core.chat_history import BaseChatMessageHistory\n", + "from pydantic import BaseModel, Field\n", + "\n", + "class ConversationSummaryMessageHistory(BaseChatMessageHistory, BaseModel):\n", + " messages: list[BaseMessage] = Field(default_factory=list)\n", + " llm: ChatOpenAI = Field(default_factory=ChatOpenAI)\n", + "\n", + " def __init__(self, llm: ChatOpenAI):\n", + " super().__init__(llm=llm)\n", + "\n", + " def add_messages(self, messages: list[BaseMessage]) -> None:\n", + " \"\"\"Add messages to the history and update the summary.\"\"\"\n", + " self.messages.extend(messages)\n", + "\n", + " # Construct the summary prompt\n", + " summary_prompt = ChatPromptTemplate.from_messages([\n", + " SystemMessagePromptTemplate.from_template(\n", + " \"Given the existing conversation summary and the new messages, \"\n", + " \"generate a new summary of the conversation. Ensure to maintain \"\n", + " \"as much relevant information as possible.\"\n", + " ),\n", + " HumanMessagePromptTemplate.from_template(\n", + " \"Existing conversation summary:\\n{existing_summary}\\n\\n\"\n", + " \"New messages:\\n{messages}\"\n", + " )\n", + " ])\n", + "\n", + " # Format the messages and invoke the LLM\n", + " new_summary = self.llm.invoke(\n", + " summary_prompt.format_messages(\n", + " existing_summary=self.messages,\n", + " messages=messages\n", + " )\n", + " )\n", + "\n", + " # Replace the existing history with a single system summary message\n", + " self.messages = [SystemMessage(content=new_summary.content)]\n", + "\n", + " def clear(self) -> None:\n", + " \"\"\"Clear the history.\"\"\"\n", + " self.messages = []\n", + "\n", + "# Create get_summary_chat_history function for summary memory\n", + "summary_chat_map = {}\n", + "\n", + "def get_summary_chat_history(session_id: str, llm: ChatOpenAI) -> ConversationSummaryMessageHistory:\n", + " if session_id not in summary_chat_map:\n", + " summary_chat_map[session_id] = ConversationSummaryMessageHistory(llm=llm)\n", + " return summary_chat_map[session_id]" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "0U4aYuOeL9cx" + }, + "source": [ + "### Memory type #3: Window Buffer Memory - Keep Latest Interactions\n", + "\n", + "An LCEL-compatible alternative to `ConversationBufferWindowMemory`. Window memory where we keep only the last k interactions in our memory and intentionally drop the oldest ones" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "metadata": { + "id": "vJ5-LUOlL9cx" + }, + "outputs": [], + "source": [ + "class BufferWindowMessageHistory(BaseChatMessageHistory, BaseModel):\n", + " messages: list[BaseMessage] = Field(default_factory=list)\n", + " k: int = Field(default_factory=int)\n", + "\n", + " def __init__(self, k: int):\n", + " super().__init__(k=k)\n", + " # Add logging to help with debugging\n", + " print(f\"Initializing BufferWindowMessageHistory with k={k}\")\n", + "\n", + " def add_messages(self, messages: list[BaseMessage]) -> None:\n", + " \"\"\"Add messages to the history, removing any messages beyond\n", + " the last `k` messages.\n", + " \"\"\"\n", + " self.messages.extend(messages)\n", + " # Add logging to help with debugging\n", + " if len(self.messages) > self.k:\n", + " print(f\"Truncating history from {len(self.messages)} to {self.k} messages\")\n", + " self.messages = self.messages[-self.k:]\n", + "\n", + " def clear(self) -> None:\n", + " \"\"\"Clear the history.\"\"\"\n", + " self.messages = []\n", + "\n", + "# Create get_chat_history function for window memory\n", + "window_chat_map = {}\n", + "\n", + "def get_window_chat_history(session_id: str, k: int = 4) -> BufferWindowMessageHistory:\n", + " print(f\"get_window_chat_history called with session_id={session_id} and k={k}\")\n", + " if session_id not in window_chat_map:\n", + " window_chat_map[session_id] = BufferWindowMessageHistory(k=k)\n", + " return window_chat_map[session_id]" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "QyU-lr-2L9cx" + }, + "source": [ + "### Memory type #4: Window + Summary Hybrid\n", + "\n", + "An LCEL-compatible alternative to `ConversationSummaryBufferMemory`. Combines the benefits of both summary and buffer window memory." + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "metadata": { + "id": "2vBeBpE1L9cx" + }, + "outputs": [], + "source": [ + "class ConversationSummaryBufferMessageHistory(BaseChatMessageHistory, BaseModel):\n", + " messages: list[BaseMessage] = Field(default_factory=list)\n", + " llm: ChatOpenAI = Field(default_factory=ChatOpenAI)\n", + " k: int = Field(default_factory=int)\n", + "\n", + " def __init__(self, llm: ChatOpenAI, k: int):\n", + " super().__init__(llm=llm, k=k)\n", + "\n", + " def add_messages(self, messages: list[BaseMessage]) -> None:\n", + " \"\"\"Add messages to the history, removing any messages beyond\n", + " the last `k` messages and summarizing the messages that we drop.\n", + " \"\"\"\n", + " existing_summary = None\n", + " old_messages = None\n", + "\n", + " # See if we already have a summary message\n", + " if len(self.messages) > 0 and isinstance(self.messages[0], SystemMessage):\n", + " existing_summary = self.messages.pop(0)\n", + "\n", + " # Add the new messages to the history\n", + " self.messages.extend(messages)\n", + "\n", + " # Check if we have too many messages\n", + " if len(self.messages) > self.k:\n", + " # Pull out the oldest messages...\n", + " old_messages = self.messages[:-self.k]\n", + " # ...and keep only the most recent messages\n", + " self.messages = self.messages[-self.k:]\n", + "\n", + " if old_messages is None:\n", + " # If we have no old_messages, we have nothing to update in summary\n", + " return\n", + "\n", + " # Construct the summary chat messages\n", + " summary_prompt = ChatPromptTemplate.from_messages([\n", + " SystemMessagePromptTemplate.from_template(\n", + " \"Given the existing conversation summary and the new messages, \"\n", + " \"generate a new summary of the conversation. Ensure to maintain \"\n", + " \"as much relevant information as possible.\"\n", + " ),\n", + " HumanMessagePromptTemplate.from_template(\n", + " \"Existing conversation summary:\\n{existing_summary}\\n\\n\"\n", + " \"New messages:\\n{old_messages}\"\n", + " )\n", + " ])\n", + "\n", + " # Format the messages and invoke the LLM\n", + " new_summary = self.llm.invoke(\n", + " summary_prompt.format_messages(\n", + " existing_summary=existing_summary or \"No previous summary\",\n", + " old_messages=old_messages\n", + " )\n", + " )\n", + "\n", + " # Prepend the new summary to the history\n", + " self.messages = [SystemMessage(content=new_summary.content)] + self.messages\n", + "\n", + " def clear(self) -> None:\n", + " \"\"\"Clear the history.\"\"\"\n", + " self.messages = []\n", + "\n", + "# Create get_chat_history function for summary buffer memory\n", + "summary_buffer_chat_map = {}\n", + "\n", + "def get_summary_buffer_chat_history(session_id: str, llm: ChatOpenAI, k: int = 4) -> ConversationSummaryBufferMessageHistory:\n", + " if session_id not in summary_buffer_chat_map:\n", + " summary_buffer_chat_map[session_id] = ConversationSummaryBufferMessageHistory(llm=llm, k=k)\n", + " return summary_buffer_chat_map[session_id]" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "xVKm61B8L9cx" + }, + "source": [ + "## Create Conversation Chains and Conversation Function" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "Xprpc4qGC0aS" + }, + "source": [ + "Create set of conversation chains that we'll be using:" + ] + }, + { + "cell_type": "code", + "execution_count": 13, + "metadata": { + "id": "I53dXttjDCgA" + }, + "outputs": [], + "source": [ + "from langchain_core.runnables.history import RunnableWithMessageHistory\n", + "from langchain_core.runnables import ConfigurableFieldSpec\n", + "\n", + "# First, create separate history factory functions for each k value\n", + "def get_window_chat_history_k6(session_id: str, k: int = 6):\n", + " return BufferWindowMessageHistory(k=k) # Changed from WindowChatHistory\n", + "\n", + "def get_window_chat_history_k12(session_id: str, k: int = 12):\n", + " return BufferWindowMessageHistory(k=k) # Changed from WindowChatHistory\n", + "\n", + "def get_summary_buffer_chat_history_k6(session_id: str, llm: ChatOpenAI, k: int = 6):\n", + " return ConversationSummaryBufferMessageHistory(llm=llm, k=k) # Changed from SummaryBufferChatHistory\n", + "\n", + "def get_summary_buffer_chat_history_k12(session_id: str, llm: ChatOpenAI, k: int = 12):\n", + " return ConversationSummaryBufferMessageHistory(llm=llm, k=k) # Changed from SummaryBufferChatHistory\n", + "\n", + "# Then update the conversation_chains dictionary to use these specific functions\n", + "conversation_chains = {\n", + " 'RunnableWithMessageHistory': RunnableWithMessageHistory(\n", + " pipeline,\n", + " get_session_history=get_chat_history,\n", + " input_messages_key=\"query\",\n", + " history_messages_key=\"history\"\n", + " ),\n", + " 'ConversationSummaryMessageHistory': RunnableWithMessageHistory(\n", + " pipeline,\n", + " get_session_history=get_summary_chat_history,\n", + " input_messages_key=\"query\",\n", + " history_messages_key=\"history\",\n", + " history_factory_config=[\n", + " ConfigurableFieldSpec(\n", + " id=\"session_id\",\n", + " annotation=str,\n", + " name=\"Session ID\",\n", + " description=\"The session ID to use for the chat history\",\n", + " default=\"summary_history\",\n", + " ),\n", + " ConfigurableFieldSpec(\n", + " id=\"llm\",\n", + " annotation=ChatOpenAI,\n", + " name=\"LLM\",\n", + " description=\"The LLM to use for the conversation summary\",\n", + " default=llm,\n", + " )\n", + " ]\n", + " ),\n", + " 'BufferWindowMessageHistory(k=6)': RunnableWithMessageHistory(\n", + " pipeline,\n", + " get_session_history=get_window_chat_history_k6, # Changed to k6 specific function\n", + " input_messages_key=\"query\",\n", + " history_messages_key=\"history\",\n", + " history_factory_config=[\n", + " ConfigurableFieldSpec(\n", + " id=\"session_id\",\n", + " annotation=str,\n", + " name=\"Session ID\",\n", + " description=\"The session ID to use for the chat history\",\n", + " default=\"window_history_k6\",\n", + " ),\n", + " ConfigurableFieldSpec(\n", + " id=\"k\",\n", + " annotation=int,\n", + " name=\"k\",\n", + " description=\"The number of messages to keep in the history\",\n", + " default=6,\n", + " )\n", + " ]\n", + " ),\n", + " 'BufferWindowMessageHistory(k=12)': RunnableWithMessageHistory(\n", + " pipeline,\n", + " get_session_history=get_window_chat_history_k12, # Changed to k12 specific function\n", + " input_messages_key=\"query\",\n", + " history_messages_key=\"history\",\n", + " history_factory_config=[\n", + " ConfigurableFieldSpec(\n", + " id=\"session_id\",\n", + " annotation=str,\n", + " name=\"Session ID\",\n", + " description=\"The session ID to use for the chat history\",\n", + " default=\"window_history_k12\",\n", + " ),\n", + " ConfigurableFieldSpec(\n", + " id=\"k\",\n", + " annotation=int,\n", + " name=\"k\",\n", + " description=\"The number of messages to keep in the history\",\n", + " default=12,\n", + " )\n", + " ]\n", + " ),\n", + " 'ConversationSummaryBufferMessageHistory(k=6)': RunnableWithMessageHistory(\n", + " pipeline,\n", + " get_session_history=get_summary_buffer_chat_history_k6, # Changed to k6 specific function\n", + " input_messages_key=\"query\",\n", + " history_messages_key=\"history\",\n", + " history_factory_config=[\n", + " ConfigurableFieldSpec(\n", + " id=\"session_id\",\n", + " annotation=str,\n", + " name=\"Session ID\",\n", + " description=\"The session ID to use for the chat history\",\n", + " default=\"summary_buffer_k6\",\n", + " ),\n", + " ConfigurableFieldSpec(\n", + " id=\"llm\",\n", + " annotation=ChatOpenAI,\n", + " name=\"LLM\",\n", + " description=\"The LLM to use for the conversation summary\",\n", + " default=llm,\n", + " ),\n", + " ConfigurableFieldSpec(\n", + " id=\"k\",\n", + " annotation=int,\n", + " name=\"k\",\n", + " description=\"The number of messages to keep in the history\",\n", + " default=6,\n", + " )\n", + " ]\n", + " ),\n", + " 'ConversationSummaryBufferMessageHistory(k=12)': RunnableWithMessageHistory(\n", + " pipeline,\n", + " get_session_history=get_summary_buffer_chat_history_k12, # Changed to k12 specific function\n", + " input_messages_key=\"query\",\n", + " history_messages_key=\"history\",\n", + " history_factory_config=[\n", + " ConfigurableFieldSpec(\n", + " id=\"session_id\",\n", + " annotation=str,\n", + " name=\"Session ID\",\n", + " description=\"The session ID to use for the chat history\",\n", + " default=\"summary_buffer_k12\",\n", + " ),\n", + " ConfigurableFieldSpec(\n", + " id=\"llm\",\n", + " annotation=ChatOpenAI,\n", + " name=\"LLM\",\n", + " description=\"The LLM to use for the conversation summary\",\n", + " default=llm,\n", + " ),\n", + " ConfigurableFieldSpec(\n", + " id=\"k\",\n", + " annotation=int,\n", + " name=\"k\",\n", + " description=\"The number of messages to keep in the history\",\n", + " default=12,\n", + " )\n", + " ]\n", + " )\n", + "}" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "XSB2VmaNL9cx" + }, + "source": [ + "Let's define the conversation function:" + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "metadata": { + "id": "YDWF27jvL9cx" + }, + "outputs": [], + "source": [ + "from tqdm.auto import tqdm\n", + "import openai\n", + "\n", "queries = [\n", " \"Good morning AI?\",\n", " \"\"\"My interest here is to explore the potential of integrating Large\n", @@ -243,7 +709,7 @@ " \"\"\"So the retrieved information is given to the chatbot / QA system as plain\n", " text? But then how do we pass in the original query? How can the system\n", " distinguish between a user's query and all of this additional information?\"\"\",\n", - " \"\"\"That doesn't seem correct to me, my question is \u2014 if we are giving the\n", + " \"\"\"That doesn't seem correct to me, my question is — if we are giving the\n", " chatbot / QA system the user's query AND retrieved information from an\n", " external knowledge base, and it's all fed into the model as plain text,\n", " how does the model know what part of the plain text is a query vs. retrieved\n", @@ -258,195 +724,134 @@ " # we loop through the conversation above, counting token usage as we go\n", " for user_query in tqdm(queries):\n", " try:\n", - " res = count_tokens(conversation_chain, user_query)\n", + " # Get the history factory function name\n", + " history_factory = conversation_chain.get_session_history.__name__\n", + "\n", + " # Create appropriate config based on history factory type\n", + " if history_factory == \"get_summary_chat_history\":\n", + " config = {\"configurable\": {\"session_id\": \"summary_history\", \"llm\": llm}}\n", + " elif history_factory in [\"get_window_chat_history_k6\", \"get_window_chat_history_k12\"]:\n", + " k = 6 if \"k6\" in history_factory else 12\n", + " config = {\"configurable\": {\"session_id\": f\"window_history_k{k}\", \"k\": k}}\n", + " elif history_factory in [\"get_summary_buffer_chat_history_k6\", \"get_summary_buffer_chat_history_k12\"]:\n", + " k = 6 if \"k6\" in history_factory else 12\n", + " config = {\"configurable\": {\"session_id\": f\"summary_buffer_k{k}\", \"llm\": llm, \"k\": k}}\n", + " else:\n", + " config = {\"configurable\": {\"session_id\": \"basic_history\"}}\n", + "\n", + " res = count_tokens(\n", + " conversation_chain,\n", + " user_query,\n", + " config=config\n", + " )\n", " tokens_used.append(res['token_count'])\n", - " except openai.error.InvalidRequestError:\n", - " # we hit the token limit of the model, so break\n", + " except (openai.APIError, openai.APIConnectionError, openai.RateLimitError, openai.APIStatusError) as e:\n", + " # we hit the token limit of the model or another API error, so break\n", + " print(f\"Hit error: {e}\")\n", " break\n", " return tokens_used" ] }, { - "attachments": {}, "cell_type": "markdown", "metadata": { - "id": "Xprpc4qGC0aS" + "id": "5DcHVsn2L9cy" }, "source": [ - "Create set of conversation chains that we'll be using:" + "## Run" ] }, { "cell_type": "code", - "execution_count": 7, - "metadata": { - "id": "I53dXttjDCgA" - }, - "outputs": [], - "source": [ - "conversation_chains = {\n", - " 'ConversationBufferMemory': ConversationChain(\n", - " llm=llm, memory=ConversationBufferMemory()\n", - " ),\n", - " 'ConversationSummaryMemory': ConversationChain(\n", - " llm=llm, memory=ConversationSummaryMemory(llm=llm)\n", - " ),\n", - " 'ConversationBufferWindowMemory(k=6)': ConversationChain(\n", - " llm=llm, memory=ConversationBufferWindowMemory(k=6)\n", - " ),\n", - " 'ConversationBufferWindowMemory(k=12)': ConversationChain(\n", - " llm=llm, memory=ConversationBufferWindowMemory(k=12)\n", - " ),\n", - " 'ConversationSummaryBufferMemory(k=6)': ConversationChain(\n", - " llm=llm, memory=ConversationSummaryBufferMemory(\n", - " llm=llm,\n", - " max_token_limit=650\n", - " )\n", - " ),\n", - " 'ConversationSummaryBufferMemory(k=12)': ConversationChain(\n", - " llm=llm, memory=ConversationSummaryBufferMemory(\n", - " llm=llm,\n", - " max_token_limit=1_300\n", - " )\n", - " )\n", - "}" - ] - }, - { - "cell_type": "code", - "execution_count": 8, + "execution_count": 17, "metadata": { "colab": { "base_uri": "https://localhost:8080/", - "height": 441, + "height": 1000, "referenced_widgets": [ - "2be4d4a7e596437db88f541b8d34bc53", - "36525328bdf1428fae13136f8a0bc7bb", - "bf6a1f2dc2c64d5991e9068e84b650b0", - "0dbec763335c48fcb2e2961fc29fc1b8", - "b88a4fd2be2f493190aeb316912df69b", - "4d9d241c61c14e37832d48e97f5b645a", - "7d017ac6d0874d95a3947746f68d6108", - "7743eca3ee7342ad86537e0c16e29b05", - "5c5d01ecf397463c939e78951b96a7c5", - "70d01f9e39ab4d84a7ca80971033ef5c", - "367c66d5b6014e0db51334e6ed05becc", - "b177914703a847479f2e5b876dbdf4f7", - "ded0b567e5f944cf9e817ae7b86fe93c", - "8708bdb3db13488bb097736bd4d5f316", - "026aaf0f075f4d44b1fa22a91274e975", - "47a83c8c9f4c4ba0adf125421f213d70", - "7645bb2b57804bcca7d667267ac33bd0", - "885c15fafa85427dac68dc0a38c7a896", - "aaabd7206abd43ed885fb6cde41375c5", - "2d3db1292fa14d6780d79eea806f3541", - "baa29f094aec4c16868f6f9affc241cd", - "39e5c5d2fc574d31919310ea8c1c9fc8", - "2cd828e3928f45cc8cf6e0050cd4f6a7", - "04daff3502b947c7887ac4d0f56fd70a", - "b9d86a3e1fe549bc93129f9235552482", - "02add01c37654a538f13010d7c71e753", - "c154bffb05d740709b59a5921b4a2c3d", - "ab269d223f0a4f128b2bf565dc4aa02c", - "42467f81fade4b58b59b0576b5312d64", - "acdb823e0b0c4a169526205fb6416ac8", - "8329f7d3193441f9bd9e1b2f2cda1dcc", - "6f4a826a519649c7842697dc52a02b3a", - "6493e6f0939f41eb8dea255ca944c57a", - "86387f7978d24544bc10cae2c488928a", - "d9c9b754057946f8b18f686366fac93f", - "b082e24b6df0495f949aa8d03d0dc519", - "0631309faf8847159f4dc15618fedab6", - "187cb0cc5b0849b3af5a54a5ad8393cd", - "d684d579ad0b45c189bbb9de17a59a81", - "c25f23a018044eb0b074ba8faec9ebb5", - "cb98068b72274befb65d34ce6ded8b15", - "356eccd305c44c92b3a768a7a497cd5e", - "1ac2a1b16ad241b993bcf97955e81577", - "35c22ad0b9074e408d58fcac08605e85", - "4455e527a2744efc88c42f1cbec1536e", - "0d047388c6b5438188a16a6cfeefbf23", - "be219119da0d484f92931324f6746f95", - "031b1285627e4865bad1199ee26906f1", - "3cc548e94bfd4c96b8ede3c713a6ca2e", - "129cbb406d9645cd814c34aac2b647ee", - "c9a8a0c6aea54aa79c2de4ff3ccf77e8", - "67cda98e1f2a48d5a8f3b54ff4f8f64f", - "fa56f77bd85b4e459ae8cb86759164eb", - "8865b266def842c181b6245f0e91b26c", - "cd623dfbc84045eea812063c7a0c1dcc", - "0cb7de9161b2409aa38b1860eb8cd990", - "e30e6c6d98c94f4aa86c6161e12b3c98", - "7ebf97578b32483b91b69d0d05ffa5a2", - "202d247919764f65b361bcef75f5642a", - "c361fca427d44e8cb0271fe2b6f4302e", - "5e808dbcc14946538b1584192e483e51", - "2707427a681a441cb4defca3d34bb748", - "9f2d7cead4cc434ea935a5ec1cf1e47b", - "2522cf65c1ba426ba27f030acc0633d9", - "a772967ef93844d0b00f027ca4134c19", - "a02cc655fd6a4e3095f2aa711a2c9a13", - "79826df9c5cc404f95d4ce8ebcfa20ff", - "30ff2b3d2b2448f3b0f68d4714684db3", - "85b05228efb54ab5a5395096bc1e699e", - "4b6924fee3fd412c8ee406fba3e6d4b4", - "f6e785ee417141cb803dfd7f03254f79", - "9a4d6fa52f6f474b9f8de85e4a8b3edf", - "0436cae6211f4cd8a2ed5990f02cbdb8", - "495465f82dc242f49b4c146b6458e5e4", - "ed9cc4b3cfc04062b38047a367ba528b", - "d08875d9c0744100ad7c3ced1aeeb2e3", - "d3ec6ba8d29f41bbaf62f07a138226a7", - "9dc25233e04e4b8cb3a6fa79f7eb1e23", - "816e625042194436af8a193436f51e26", - "b367539fab28453698f549eb286c98f8", - "84b54b8074a04e6e85540d4faf5e27d5", - "20dcc9c584dd47789c0b9297b4973b18", - "6d5bc0a26699408cb9dbeaf831c9949d", - "ca96a4cad9d945ccba9f1a9bf30a593d", - "8b3875c1e821407c8118dba150838b31", - "a65ecfbfb7e34465978fe2799e70319a", - "6bd35b95ef114e4b8b57fc49757381a8", - "87d981713a024b18b65ecb1c04b26bf2", - "2ff386b5a15b43f4b75ba4c6fa6959a7", - "76df0865ad634f7a8061e96d8f5b80c7", - "ed889a0e039646afb9ac267b32507df3", - "f7984859eacc41c89ef33526efa1df0b", - "9e19e310ce044564842537965ee66841", - "6c13c9ebeb474c66907262c9573311d0", - "aeeb99e1086840c8b7d0807ccd0b057b", - "c0cf94856df84c12987b54b5608c258d", - "6de6c6d6f201416d850455f96b94407f", - "6c883833b99448fba7ff6c8b2cc30aee", - "3169a365eb504f838c6726a93adf1a93", - "6193595b5824404b8e6c9428612d67ee", - "c423831fe1e543458ab1fc16d68dbc42", - "79b582179bea4d299eaeabb6b54ad49b", - "073d4b875d57463398fc936f9272ffa1", - "a79b233bb44a4e699f7241a0bee9fe42", - "8e45498c5bc748479c8dcd1e615e8add", - "aa288fc687404376ab55bdf5dd036ce1", - "fd2a27dff42f43b3b9c25d678d6c32f0", - "470487e6768148049b091513538236bb", - "e9443b0de3d34fe5b9f51b2a419acff0", - "c49af1d9b28741299e25dd2646099e23" + "daab3b49ad8740d59df0994db74a57a3", + "e8b3fa88b68543119c53d889d528a999", + "65acb879ec39416cbd76b308baab2114", + "6c01285eea9c4f03bedf7d55a6f49252", + "db072b06b6aa4e1595efb15f4a7e78c5", + "2e5184aa569a4504ac89bfc00988de9b", + "ef2d497fe30343199e068bc125fc9b3e", + "40a2dc36f66542f4b40aac039b20cfd2", + "c0bb84b9215c4d109dfe530a11d9daba", + "2b2ee47b98f94db2810c72552022c1c0", + "9e73c521564d4bab95c77e21ef4e3f0f", + "32ba0d9d4d89486ba2376f1b37019145", + "b4b12096937a4b7c83e83ff3db6c447b", + "a8344f58cff54c4db313c693a46b7ddc", + "61044e0b34fe40f7a82dd028d351a323", + "0131fb1aa6d04ac99d716fd865b50de4", + "511dd092c39b499ea08e2114b6aaee34", + "7c569deaafca4641a8636a1a069efcc4", + "f77a980129664559adc026b217143cc1", + "26fe33dce8f94b249ac45e95cda1b918", + "8a54a184c7a84a5eb6206421d914e82b", + "baa72e62064e44139cfff25cc88feac9", + "f639f24ecac54cb4bffa10753f85e88c", + "6414efabbaf0408a9d9b6715e39de480", + "e0eb7dffc9b04cb7a2c07c73493c6179", + "86abfb3b370f4ff1bd54b38ce105d57a", + "c9f458f68b2b4d9e9c6d948772e55fee", + "45121c4b6ec4429ebc4a15362a2aabb8", + "9c00596b9f8d47c7b015d91a5e939984", + "d1282a8fbe194b95ac15fa33f948507d", + "88a33ed068994fce9617e673f3ad1914", + "0cfce9c5d4bf4d928afb6a5aee181c22", + "1aebe25b3dd54f4e9b2c10bb93e66823", + "63496596987f4794859adc454a43f7eb", + "8025621198f54c9aa1f5af80f038bb79", + "d5ea0e738e2148aea413dbf223f591e1", + "17887dc8f64b472db15c20d7c629b20e", + "1e8dc4c3c8e540f8b3318dde53ae8870", + "f15aa248864f41d0a8b914f523e1a3b8", + "150cd8937c1e4db1a6602eac95426c51", + "a48079c8409f44bdabf36c7a15922e3f", + "ea0af4aac9d74dd78714f185170d509e", + "ee820d9d924e45e9bb2343a60c5f56e6", + "be098632bf264ef58f540275d98ac497", + "481d6a256c394b7693f7640c01e0ec09", + "a23e829a4cc54c94810d103313518d1a", + "e48ab411aa5c418486956af879420669", + "8660d8ba8ef541f8be529e50dba8f58f", + "77986febcec7449da7f4faf4e2573097", + "5b1c37edce524ff0b12640ea59940211", + "ca33552c6e494b78a245dd826706fc7e", + "1dd49d5a7a64474fb7f2d148126915c8", + "268c2c10b8a146838afcaf31e9e57a90", + "db78179806d84fd9955726e74cd1d68d", + "79a6af3b3a89475ca2c5ec2853179d17", + "e33a510efbaa4a02a946600142c78e0b", + "32ab58dd34664537aeeead0ee1631dd2", + "5b985f564cc3435fb9922ceef83546f8", + "d28b3a34f0e7468898454dbbf8e80f1b", + "36a79ef2b37343369e72fc6154a53e7d", + "d18abf8c84a3415dafc1e4515bf36e7e", + "5ea97f09691240d3867f7dfa64f2f1a8", + "1d600b2fb50b48ba90881b9b4923198f", + "d64e8544c0094c04853247fa83a42f73", + "0e3c92387b3c433db3bf285e840726a0", + "aa24a99fc650485f8230a8a673246014" ] }, "id": "Rxirb9Vt5oOj", - "outputId": "1a29133b-e7f2-4ac3-a80a-2dd674e3d8c7" + "outputId": "4a7b2c18-81be-4c0d-ec53-c7094695a577" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ - "ConversationBufferMemory\n" + "RunnableWithMessageHistory\n" ] }, { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "2be4d4a7e596437db88f541b8d34bc53", + "model_id": "daab3b49ad8740d59df0994db74a57a3", "version_major": 2, "version_minor": 0 }, @@ -461,13 +866,40 @@ "name": "stdout", "output_type": "stream", "text": [ - "ConversationSummaryMemory\n" + "Spent a total of 74 tokens\n", + "Spent a total of 629 tokens\n", + "Spent a total of 1671 tokens\n", + "Spent a total of 2429 tokens\n", + "Spent a total of 3476 tokens\n", + "Spent a total of 4186 tokens\n", + "Spent a total of 5057 tokens\n", + "Spent a total of 5765 tokens\n", + "Spent a total of 6460 tokens\n", + "Spent a total of 7142 tokens\n", + "Spent a total of 7726 tokens\n", + "Spent a total of 8516 tokens\n", + "Spent a total of 9216 tokens\n", + "Spent a total of 10228 tokens\n", + "Spent a total of 11128 tokens\n", + "Spent a total of 12060 tokens\n", + "Spent a total of 13096 tokens\n", + "Spent a total of 13743 tokens\n", + "Spent a total of 14353 tokens\n", + "Spent a total of 15006 tokens\n", + "Spent a total of 15892 tokens\n", + "Spent a total of 16559 tokens\n", + "Spent a total of 17330 tokens\n", + "Spent a total of 18028 tokens\n", + "Spent a total of 18610 tokens\n", + "Spent a total of 19295 tokens\n", + "Spent a total of 19813 tokens\n", + "ConversationSummaryMessageHistory\n" ] }, { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "b177914703a847479f2e5b876dbdf4f7", + "model_id": "32ba0d9d4d89486ba2376f1b37019145", "version_major": 2, "version_minor": 0 }, @@ -482,13 +914,40 @@ "name": "stdout", "output_type": "stream", "text": [ - "ConversationBufferWindowMemory(k=6)\n" + "Spent a total of 225 tokens\n", + "Spent a total of 1643 tokens\n", + "Spent a total of 3250 tokens\n", + "Spent a total of 3180 tokens\n", + "Spent a total of 4582 tokens\n", + "Spent a total of 3566 tokens\n", + "Spent a total of 3500 tokens\n", + "Spent a total of 2455 tokens\n", + "Spent a total of 2196 tokens\n", + "Spent a total of 1950 tokens\n", + "Spent a total of 1338 tokens\n", + "Spent a total of 2458 tokens\n", + "Spent a total of 2732 tokens\n", + "Spent a total of 4310 tokens\n", + "Spent a total of 4180 tokens\n", + "Spent a total of 4127 tokens\n", + "Spent a total of 4190 tokens\n", + "Spent a total of 2490 tokens\n", + "Spent a total of 3024 tokens\n", + "Spent a total of 4575 tokens\n", + "Spent a total of 4727 tokens\n", + "Spent a total of 3096 tokens\n", + "Spent a total of 2633 tokens\n", + "Spent a total of 2523 tokens\n", + "Spent a total of 2492 tokens\n", + "Spent a total of 2615 tokens\n", + "Spent a total of 2336 tokens\n", + "BufferWindowMessageHistory(k=6)\n" ] }, { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "2cd828e3928f45cc8cf6e0050cd4f6a7", + "model_id": "f639f24ecac54cb4bffa10753f85e88c", "version_major": 2, "version_minor": 0 }, @@ -503,13 +962,67 @@ "name": "stdout", "output_type": "stream", "text": [ - "ConversationBufferWindowMemory(k=12)\n" + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 74 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 511 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 147 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 468 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 601 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 542 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 888 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 310 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 599 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 401 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 320 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 691 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 527 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 785 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 760 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 390 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 630 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 127 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 586 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 595 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 657 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 501 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 552 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 435 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 529 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 508 tokens\n", + "Initializing BufferWindowMessageHistory with k=6\n", + "Spent a total of 583 tokens\n", + "BufferWindowMessageHistory(k=12)\n" ] }, { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "86387f7978d24544bc10cae2c488928a", + "model_id": "63496596987f4794859adc454a43f7eb", "version_major": 2, "version_minor": 0 }, @@ -524,13 +1037,67 @@ "name": "stdout", "output_type": "stream", "text": [ - "ConversationSummaryBufferMemory(k=6)\n" + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 74 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 461 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 148 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 506 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 717 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 548 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 952 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 501 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 582 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 403 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 303 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 586 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 566 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 663 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 748 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 389 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 681 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 133 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 601 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 495 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 641 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 543 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 529 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 426 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 425 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 547 tokens\n", + "Initializing BufferWindowMessageHistory with k=12\n", + "Spent a total of 602 tokens\n", + "ConversationSummaryBufferMessageHistory(k=6)\n" ] }, { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "4455e527a2744efc88c42f1cbec1536e", + "model_id": "481d6a256c394b7693f7640c01e0ec09", "version_major": 2, "version_minor": 0 }, @@ -542,56 +1109,48 @@ "output_type": "display_data" }, { - "data": { - "application/vnd.jupyter.widget-view+json": { - "model_id": "0cb7de9161b2409aa38b1860eb8cd990", - "version_major": 2, - "version_minor": 0 - }, - "text/plain": [ - "Downloading (\u2026)olve/main/vocab.json: 0%| | 0.00/1.04M [00:00" + "
" ] }, - "metadata": { - "needs_background": "light" - }, + "metadata": {}, "output_type": "display_data" } ], "source": [ - "import seaborn as sns\n", "import matplotlib.pyplot as plt\n", + "import seaborn as sns\n", + "\n", + "plt.figure(figsize=(12,8))\n", + "max_tokens = 4096\n", + "\n", + "colors = [\"#1c17ff\", \"#738FAB\", \"#f77f00\", \"#fcbf49\", \"#38c172\", \"#4dc0b5\"]\n", "\n", + "for i, (key, count) in enumerate(counts.items()):\n", + " color = colors[i]\n", + " sns.lineplot(\n", + " x=range(1, len(count)+1),\n", + " y=count,\n", + " label=key,\n", + " color=color\n", + " )\n", + " if max_tokens in count:\n", + " plt.plot(\n", + " len(count), max_tokens, marker=\"X\", color=\"red\", markersize=10\n", + " )\n", + "\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "r19MnfM3L9cy" + }, + "source": [ + "Or, alternatively, a logarithmic plot to show the non `RunnableWithMessageHistory` (blue line) plots more clearly." + ] + }, + { + "cell_type": "code", + "execution_count": 21, + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 676 + }, + "id": "d5vptF9SL9cy", + "outputId": "a93093a8-c303-4b93-813b-fa3a494126e0" + }, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ "plt.figure(figsize=(12,8))\n", "max_tokens = 4096\n", "\n", @@ -674,11 +1308,11 @@ " len(count), max_tokens, marker=\"X\", color=\"red\", markersize=10\n", " )\n", "\n", + "plt.yscale('log') # Set y-axis to logarithmic scale\n", "plt.show()" ] }, { - "attachments": {}, "cell_type": "markdown", "metadata": { "id": "jZuSLqremRX0" @@ -693,13 +1327,23 @@ "provenance": [] }, "kernelspec": { - "display_name": "Python 3", + "display_name": "pinecone1", + "language": "python", "name": "python3" }, "language_info": { - "name": "python" + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.11.4" } }, "nbformat": 4, "nbformat_minor": 0 -} \ No newline at end of file +}