DOCSP-49240: Add LangChain semantic cache and Local RAG notebooks

ai-integrations/langchain-graphrag.ipynb

@@ -5,7 +5,7 @@
    "id": "b5dcbf95-9a30-416d-afed-d5b2bf0e8651",
    "metadata": {},
    "source": [
-    "# GraphRAG with MongoDB and LangChain\n",
+    "# LangChain MongoDB Integration - GraphRAG\n",
     "\n",
     "This notebook is a companion to the [GraphRAG with MongoDB and LangChain](https://www.mongodb.com/docs/atlas/atlas-vector-search/ai-integrations/langchain/graph-rag/) tutorial. Refer to the page for set-up instructions and detailed explanations.\n",
     "\n",

ai-integrations/langchain-hybrid-search.ipynb

@@ -4,7 +4,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Atlas Vector Search - LangChain Integration - Hybrid Search"
+    "# LangChain MongoDB Integration - Hybrid Search"
    ]
   },
   {

ai-integrations/langchain-local-rag.ipynb

@@ -0,0 +1,251 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# LangChain MongoDB Integration - Implement RAG Locally"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "This notebook is a companion to the [LangChain Local RAG](https://www.mongodb.com/docs/atlas/atlas-vector-search/ai-integrations/langchain/get-started/) tutorial. Refer to the page for set-up instructions and detailed explanations.\n",
+    "\n",
+    "<a target=\"_blank\" href=\"https://colab.research.google.com/github/mongodb/docs-notebooks/blob/main/ai-integrations/langchain-local-rag.ipynb\">\n",
+    "  <img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/>\n",
+    "</a>"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "vscode": {
+     "languageId": "shellscript"
+    }
+   },
+   "source": [
+    "## Create a local Atlas deployment\n",
+    "\n",
+    "Run the following commands in your terminal to set up your local Atlas deployment. \n",
+    "\n",
+    "```\n",
+    "atlas deployments setup\n",
+    "curl  https://atlas-education.s3.amazonaws.com/sampledata.archive -o sampledata.archive\n",
+    "mongorestore --archive=sampledata.archive --port=<port-number>\n",
+    "```"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "vscode": {
+     "languageId": "shellscript"
+    }
+   },
+   "source": [
+    "## Set up the environment"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "vscode": {
+     "languageId": "shellscript"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "pip install --quiet --upgrade pymongo langchain gpt4all sentence_transformers langchain-huggingface"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "MONGODB_URI = (\"mongodb://localhost:64983/?directConnection=true\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Create embeddings with local model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from pymongo import MongoClient\n",
+    "from sentence_transformers import SentenceTransformer\n",
+    "\n",
+    "# Connect to your local Atlas deployment or Atlas Cluster\n",
+    "client = MongoClient(MONGODB_URI)\n",
+    "\n",
+    "# Select the sample_airbnb.listingsAndReviews collection\n",
+    "collection = client[\"sample_airbnb\"][\"listingsAndReviews\"]\n",
+    "\n",
+    "# Load the embedding model (https://huggingface.co/mixedbread-ai/mxbai-embed-large-v1)\n",
+    "model_path = \"<path-to-save-model>\"\n",
+    "model = SentenceTransformer('mixedbread-ai/mxbai-embed-large-v1')\n",
+    "model.save(model_path)\n",
+    "model = SentenceTransformer(model_path)\n",
+    "\n",
+    "# Define function to generate embeddings\n",
+    "def get_embedding(text):\n",
+    "    return model.encode(text).tolist()\n",
+    "\n",
+    "# Filters for only documents with a summary field and without an embeddings field\n",
+    "filter = { '$and': [ { 'summary': { '$exists': True, \"$nin\": [ None, \"\" ] } }, { 'embeddings': { '$exists': False } } ] }\n",
+    "\n",
+    "# Creates embeddings for subset of the collection\n",
+    "updated_doc_count = 0\n",
+    "for document in collection.find(filter).limit(50):\n",
+    "    text = document['summary']\n",
+    "    embedding = get_embedding(text)\n",
+    "    collection.update_one({ '_id': document['_id'] }, { \"$set\": { 'embeddings': embedding } }, upsert=True)\n",
+    "    updated_doc_count += 1\n",
+    "\n",
+    "print(\"Documents updated: {}\".format(updated_doc_count))\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Configure the vector store"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain_huggingface import HuggingFaceEmbeddings\n",
+    "\n",
+    "embedding_model = HuggingFaceEmbeddings(model_name=\"mixedbread-ai/mxbai-embed-large-v1\")\n",
+    "\n",
+    "from langchain_mongodb import MongoDBAtlasVectorSearch\n",
+    "# Instantiate vector store\n",
+    "vector_store = MongoDBAtlasVectorSearch.from_connection_string(\n",
+    "   connection_string = MONGODB_URI,\n",
+    "   namespace = \"sample_airbnb.listingsAndReviews\",\n",
+    "   embedding=embedding_model,\n",
+    "   index_name=\"vector_index\",\n",
+    "   embedding_key=\"embeddings\",\n",
+    "   text_key=\"summary\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "vector_store.create_vector_search_index(\n",
+    "  dimensions = 1024,       # The dimensions of the vector embeddings to be indexed\n",
+    "  wait_until_complete = 60 # Number of seconds to wait for the index to build (can take around a minute)\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Implement RAG with a local LLM\n",
+    "Before running the following code, [download the local model](https://gpt4all.io/models/gguf/mistral-7b-openorca.gguf2.Q4_0.gguf)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler\n",
+    "from langchain_community.llms import GPT4All\n",
+    "\n",
+    "# Configure the LLM\n",
+    "local_path = \"<path-to-model>\"\n",
+    "\n",
+    "# Callbacks support token-wise streaming\n",
+    "callbacks = [StreamingStdOutCallbackHandler()]\n",
+    "\n",
+    "# Verbose is required to pass to the callback manager\n",
+    "llm = GPT4All(model=local_path, callbacks=callbacks, verbose=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain_core.prompts import PromptTemplate\n",
+    "from langchain_core.output_parsers import StrOutputParser\n",
+    "from langchain_core.runnables import RunnablePassthrough\n",
+    "import pprint\n",
+    "\n",
+    "# Instantiate Atlas Vector Search as a retriever\n",
+    "retriever = vector_store.as_retriever()\n",
+    "\n",
+    "# Define prompt template\n",
+    "template = \"\"\"\n",
+    "Use the following pieces of context to answer the question at the end.\n",
+    "{context}\n",
+    "Question: {question}\n",
+    "\"\"\"\n",
+    "custom_rag_prompt = PromptTemplate.from_template(template)\n",
+    "\n",
+    "def format_docs(docs):\n",
+    "   return \"\\n\\n\".join(doc.page_content for doc in docs)\n",
+    "\n",
+    "# Create chain   \n",
+    "rag_chain = (\n",
+    "   {\"context\": retriever | format_docs, \"question\": RunnablePassthrough()}\n",
+    "   | custom_rag_prompt\n",
+    "   | llm\n",
+    "   | StrOutputParser()\n",
+    ")\n",
+    "\n",
+    "# Prompt the chain\n",
+    "question = \"Can you recommend me a few AirBnBs that are beach houses?\"\n",
+    "answer = rag_chain.invoke(question)\n",
+    "\n",
+    "# Return source documents\n",
+    "documents = retriever.invoke(question)\n",
+    "print(\"\\nSource documents:\")\n",
+    "pprint.pprint(documents)"
+   ]
+  }
+ ],
+ "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": "3.10.12"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}