{"id":247172,"date":"2026-06-03T09:07:00","date_gmt":"2026-06-03T07:07:00","guid":{"rendered":"https:\/\/graphwise.ai\/?post_type=blog-post&p=247172"},"modified":"2026-07-10T05:33:31","modified_gmt":"2026-07-10T03:33:31","slug":"precise-semantic-retrieval-implementing-chunk-level-vector-search-with-graphdb-and-elasticsearch","status":"publish","type":"blog-post","link":"https:\/\/gws-sso-test.graphwise.ai\/de\/blog\/precise-semantic-retrieval-implementing-chunk-level-vector-search-with-graphdb-and-elasticsearch\/","title":{"rendered":"Precise Semantic Retrieval: Implementing Chunk-level Vector Search with GraphDB and Elasticsearch"},"content":{"rendered":"\n

When working with large documents, semantic search<\/a> at the document level is often insufficient. Large texts frequently contain multiple distinct topics, and indexing them as a single unit can cause truncation. This may happen either because the embeddings sent to the Graphwise Transformer client exceed 4 MB (larger documents cannot fit within gRPC limitations) or due to size restrictions of the embedding model used by the transformer.<\/p>\n\n\n\n

A more precise approach is to split documents into smaller logical segments (chunks) and generate embeddings per chunk.<\/p>\n\n\n\n

With the introduction of vector search within nested fields<\/strong> in GraphDB 11.3<\/a>, it is now possible to model and index these chunks as structured nested objects in Elasticsearch while preserving their semantic integrity. This enables accurate chunk-level vector similarity search without losing the relationship between different object fields.<\/p>\n\n\n\n

How to set it up<\/h2>\n\n\n\n

The following steps describe the complete process \u2014 from preparing chunked RDF data, through configuring the Graphwise Transformer client and GraphDB, to creating the Elasticsearch connector and executing nested vector search queries.<\/p>\n\n\n\n

Step 1: Generate chunked RDF data<\/h3>\n\n\n\n

Use the chunker.py<\/code> script to generate an RDF<\/a> dataset from a TSV file. The script can be found at https:\/\/github.com\/Ontotext-AD\/document-chunker-rdf<\/a>.<\/p>\n\n\n\n

As mentioned earlier, if the source documents are very long, Elasticsearch may index only part of the content. To avoid this behavior, split each long document into smaller chunks and index each chunk separately.<\/p>\n\n\n\n

Step 2: Start the Graphwise Transformer client and Elasticsearch<\/h3>\n\n\n\n

The Graphwise Transformer is a Python gRPC server that serves sentence embeddings via gRPC. By default, it listens on port 5050 and is started as follows:<\/p>\n\n\n\n

docker build -t graphwise-transformer:$(git rev-parse –short HEAD) .<\/strong><\/p>\n\n\n\n

docker run –rm -p 5050:5050 -e GRAPHWISE_CONFIG=\/app\/config.properties graphwise-transformer:$(git rev-parse –short HEAD)<\/strong><\/p>\n\n\n\n

This uses the default configuration for port and sentence embeddings model – port 5050<\/strong> and the sentence-transformers\/paraphrase-multilingual-MiniLM-L12-v2<\/strong> embedding model. Both can be changed by modifying the config.properties<\/code>. Further information about the client is available in the Graphwise documentation<\/a>.<\/p>\n\n\n\n

Step 3: Start GraphDB with the required properties configured<\/h3>\n\n\n\n

GraphDB can be started using the official dockerhub image or the provided distributions for Windows, Linux, and MacOS.<\/p>\n\n\n\n

Start GraphDB with the following properties configured (see the configuration reference in our documentation<\/a>):<\/p>\n\n\n\n