llm-semantic-cache/llm_semantic_cache_v1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "80JsJB4V93dw"
   },
   "source": [
    "# **LLM Serving with Apigee**\n",
    "\n",
    "<table align=\"left\">\n",
    "    <td style=\"text-align: center\">\n",
    "        <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/apigee-samples/blob/main/llm-semantic-cache/llm_semantic_cache_v1.ipynb\">\n",
    "          <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Google Colaboratory logo\\\"><br> Open in Colab\n",
    "        </a>\n",
    "      </td>\n",
    "      <td style=\"text-align: center\">\n",
    "        <a href=\"https://console.cloud.google.com/vertex-ai/colab/import/https%3A%2F%2Fraw.githubusercontent.com%2FGoogleCloudPlatform%2Fapigee-samples%2Fmain%2Fllm-semantic-cache%2Fllm_semantic_cache_v1.ipynb\">\n",
    "          <img width=\"32px\" src=\"https://lh3.googleusercontent.com/JmcxdQi-qOpctIvWKgPtrzZdJJK-J3sWE1RsfjZNwshCFgE_9fULcNpuXYTilIR2hjwN\" alt=\"Google Cloud Colab Enterprise logo\"><br> Open in Colab Enterprise\n",
    "        </a>\n",
    "      </td>    \n",
    "      <td style=\"text-align: center\">\n",
    "        <a href=\"https://console.cloud.google.com/vertex-ai/workbench/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/apigee-samples/main/llm-semantic-cache/llm_semantic_cache_v1.ipynb\">\n",
    "          <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\"><br> Open in Workbench\n",
    "        </a>\n",
    "      </td>\n",
    "      <td style=\"text-align: center\">\n",
    "        <a href=\"https://github.com/GoogleCloudPlatform/apigee-samples/blob/main/llm-semantic-cache/llm_semantic_cache_v1.ipynb\">\n",
    "          <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\"><br> View on GitHub\n",
    "        </a>\n",
    "      </td>\n",
    "</table>\n",
    "<br />\n",
    "<br />\n",
    "<br />\n",
    "\n",
    "# Semantic Cache Sample\n",
    "\n",
    "This sample performs a cache lookup of responses on Apigee's Cache layer and Vector Search as an embeddings database. It operates by comparing the vector proximity of the prompt to prior requests and using a configurable similarity score threshold.\n",
    "\n",
    "![architecture](https://github.com/GoogleCloudPlatform/apigee-samples/blob/main/llm-semantic-cache/images/arch-1.png?raw=1)\n",
    "\n",
    "# Benefits of a Semantic Cache Layer with Apigee:\n",
    "\n",
    "* **Reduced Response Times**: The cache layer significantly reduces response times for repeated queries, as Apigee efficiently stores and retrieves frequently accessed data.\n",
    "* **Improved Efficiency**: By leveraging the caching capabilities of Apigee, unnecessary calls to the underlying model will be minimized, leading to optimized LLM costs.\n",
    "* **Scalability**: The Apigee Cache Layer is managed and distributed, enhancing platform scalability without operational overhead.\n",
    "\n",
    "\n",
    "## About Vertex AI Vector Search API and Embeddings API\n",
    "\n",
    "[**Vertex AI Vector Search**](https://cloud.google.com/vertex-ai/docs/vector-search/overview) enables real-time, fast retrieval of embeddings, which powers a wide range of next-gen user experiences. It provices state-of-the-art embeddigs similarity search ([**ScaNN**](https://research.google/blog/announcing-scann-efficient-vector-similarity-search/)) that is foundational to Google services like Search, Play, and Youtube. It is a key enabler for Search applications and RAG. Vertex AI Vector Search offers speed, scale, quality and cost advantages over alternatives. It also has differentiated value added capabilities including incremental indexing, numerical and tag-based filtering, ensuring diversity of results, and auto-scaling.\n",
    "\n",
    "[**Vertex AI Embeddings API**](https://cloud.google.com/vertex-ai/generative-ai/docs/embeddings/get-text-embeddings) provides a powerful way to represent text, images, and videos as numerical vectors. It allows for finding similar content based on meaning, Suggesting relevant items based on user preferences and past interactions, classifying, clustering, and detecting outliers based on semantic relationships. It is a key component of RAG architecture allowing more natural and engaging interactions with chatbots. It supports combining information from different data types for richer insights.\n",
    "\n",
    "## How does it work?\n",
    "\n",
    "1. Prompt request is receved by an Apigee Proxy.\n",
    "2. Apigee extracts prompt contents and generates a numerical representation using the Vertex AI Embeddings API\n",
    "3. Apigee performs a semantic similarity search using Vertex AI Vector Search\n",
    "4. If there's a datapoint with a good similarity score, then perform a cache lookup using [**Apigee's Cache**](https://cloud.google.com/apigee/docs/api-platform/cache/persistence-tools#caching).\n",
    "5. If there's a cached datapoint, then return the cached LLM response, otherwise populate the Apigee Cache with the LLM respose.\n",
    "\n",
    "### Semantic cache cleanup utility\n",
    "\n",
    "This sample includes a semantic cache cleanup utility. This automation periodically cleans up stale datapoints from the Vertex AI Vector Index. Once deployed it will trigger every hour and this frequency can be adjusted by setting a new cron configuration and re-deploying the automation. Users can also trigger manual cleanup jobs. To know more about this utility, go to the following link (replace variables with your project and region settings):\n",
    "\n",
    "```\n",
    "https://console.cloud.google.com/integrations/edit/cleanup-semantic-cache-v1/locations/$REGION?project=$PROJECT\n",
    "```\n",
    "\n",
    "## Setup\n",
    "\n",
    "Use the following GCP CloudShell tutorial. Follow the instructions to deploy the sample.\n",
    "\n",
    "[![Open in Cloud Shell](https://gstatic.com/cloudssh/images/open-btn.png)](https://ssh.cloud.google.com/cloudshell/open?cloudshell_git_repo=https://github.com/GoogleCloudPlatform/apigee-samples&cloudshell_git_branch=main&cloudshell_workspace=.&cloudshell_tutorial=llm-semantic-cache/docs/cloudshell-tutorial.md)\n",
    "\n",
    "## Test Sample\n",
    "\n",
    "### Install dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "58S8GPAnHZbM"
   },
   "outputs": [],
   "source": [
    "!pip install langchain\n",
    "!pip install langchain-community\n",
    "!pip install langchain_google_vertexai\n",
    "!pip install google-cloud-aiplatform"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "KyPnkqS9Hm5I"
   },
   "source": [
    "### Authenticate your notebook environment (Colab only)\n",
    "If you are running this notebook on Google Colab, run the following cell to authenticate your environment. This step is not required if you are using Vertex AI Workbench or Colab Enterprise."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "q_-3uHjVHmA2"
   },
   "outputs": [],
   "source": [
    "import sys\n",
    "\n",
    "# Additional authentication is required for Google Colab\n",
    "if \"google.colab\" in sys.modules:\n",
    "    # Authenticate user to Google Cloud\n",
    "    from google.colab import auth\n",
    "\n",
    "    auth.authenticate_user()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "tRXV6laSVZvW"
   },
   "source": [
    "### Initialize notebook variables\n",
    "\n",
    "* PROJECT_ID: The default GCP project to use when making Vertex API calls.\n",
    "* REGION: The default location to use when making API calls.\n",
    "* API_ENDPOINT:  Desired API endpoint, e.g., apigee.iloveapimanagement.com/generate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "0WGbICuXUdCa"
   },
   "outputs": [],
   "source": [
    "from langchain_google_vertexai import VertexAI\n",
    "# Define project information\n",
    "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
    "LOCATION = \"\"  # @param {type:\"string\"}\n",
    "API_ENDPOINT = \"https://REPLACE_WITH_APIGEE_HOST/v1/samples/llm-semantic-cache\"  # @param {type:\"string\"}\n",
    "# Initialize Langchain\n",
    "model = VertexAI(\n",
    "      project=PROJECT_ID,\n",
    "      location=LOCATION,\n",
    "      api_endpoint=API_ENDPOINT,\n",
    "      api_transport=\"rest\",\n",
    "      streaming=True,\n",
    "      model_name=\"gemini-1.5-pro\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "BlmaG7qj7y33"
   },
   "source": [
    "### Test and analyze semantic cache performance\n",
    "\n",
    "This script measures and visualizes the performance of a semantic cache layer implemented using Apigee.\n",
    "\n",
    "It executes a set of prompts multiple times and records the response times for each execution.\n",
    "The script then plots the response times over the executions, highlighting the average response time."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 487
    },
    "id": "Jy-rqlcuyelr",
    "outputId": "a9c0de42-0c87-4fe6-d71c-aa4056f37577"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import time\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib as mpl\n",
    "\n",
    "\n",
    "exec = 2\n",
    "execs = []\n",
    "prompts = [\"Why is the sky blue?\",\n",
    "           \"What makes the sky blue?\",\n",
    "           \"Why does the sky is blue colored?\",\n",
    "           \"Can you explain why the sky is blue?\",\n",
    "           \"The sky is blue, why is that?\"]\n",
    "for i in range(exec):\n",
    "  for prompt in prompts:\n",
    "    start_time = time.time()\n",
    "    model.invoke(prompt)\n",
    "    response_time = time.time() - start_time\n",
    "    execs.append(response_time)\n",
    "\n",
    "mpl.rcParams['figure.figsize'] = [15, 5]\n",
    "df = pd.DataFrame(execs, columns=['Response time'])\n",
    "df['Exec'] = range(1, len(df) + 1)\n",
    "df.plot(kind='line', x='Exec', y='Response time', legend=False)\n",
    "plt.title('Semantic Cache Performance')\n",
    "plt.xlabel('Executions')\n",
    "plt.ylabel('Response Time')\n",
    "plt.xticks(df['Exec'], rotation=0)\n",
    "\n",
    "average = df['Response time'].mean()\n",
    "plt.axhline(y=average, color='r', linestyle='--', label=f'Average: {average:.2f}')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "provenance": [],
   "toc_visible": true
  },
  "kernelspec": {
   "display_name": "Python 3",
   "name": "python3"
  },
  "language_info": {
   "name": "python"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}