From 8cf8fede1db35129199cefca1847b473b5d3cb8c Mon Sep 17 00:00:00 2001 From: Kaustav Ghosh Date: Thu, 19 Sep 2024 12:28:27 +0530 Subject: [PATCH] feat: Update GraphRAG notebook with Couchbase setup instructions --- .../couchbase/GraphRAG_with_Couchbase.ipynb | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/examples_notebooks/community_contrib/couchbase/GraphRAG_with_Couchbase.ipynb b/examples_notebooks/community_contrib/couchbase/GraphRAG_with_Couchbase.ipynb index bf80577c32..e521b6ebe3 100644 --- a/examples_notebooks/community_contrib/couchbase/GraphRAG_with_Couchbase.ipynb +++ b/examples_notebooks/community_contrib/couchbase/GraphRAG_with_Couchbase.ipynb @@ -7,7 +7,21 @@ }, "source": [ "# Tutorial on GraphRAG Local Search with Couchbase Vector Store\n", - "This notebook walks through the process of setting up a search engine that combines Couchbase for storing embeddings, OpenAI's models for generating embeddings, and a local search engine for querying structured data. This is useful when you need to search through structured data using natural language queries, leveraging both machine learning and a database." + "\n", + "This notebook walks through the process of setting up a search engine that combines Couchbase for storing embeddings, OpenAI's models for generating embeddings, and a local search engine for querying structured data. \n", + "This is useful when you need to search through structured data using natural language queries, leveraging both machine learning and a database.This approach is useful for searching structured data using natural language queries, leveraging both machine learning and a database.\n", + "\n", + "## Setting up Couchbase\n", + "\n", + "Before running this notebook, set up the following in Couchbase:\n", + "\n", + "1. Create a bucket named \"graphrag-demo\" (or as specified in COUCHBASE_BUCKET_NAME)\n", + "2. Within the bucket, create a scope named \"shared\" (or as specified in COUCHBASE_SCOPE_NAME)\n", + "3. Within the scope, create a collection named \"entity_description_embeddings\" (or as specified in COUCHBASE_COLLECTION_NAME)\n", + "\n", + "These settings should match the environment variables defined in your .env file or the default values in the code.\n", + "\n", + "4. In the Couchbase Full Text Search (FTS) index section, create a new index by importing the `graphrag_demo_index.json` file. This file contains the necessary configuration for the vector search index.\n" ] }, {