OpenSearch ChatGPT Retrieval-Augmented Generation
https://opensearch.org/old_downloads.html
- Download OpenSearch 2.17.1 and unzip it
- Run OpenSearch
- Change the
config.inifile- Set your OpenSearch host and port
- Set your ElasticSearch host, port, index name, source fields to vector index, the query and the size limit
- Set your OpenAI API key in your environment variables
- Override the default OpenAI model if you want to use a different one (Optional)
- If you want to override the sentence transformer model, you must adapt the text_embedding dimension vector
- Run
webapp.py - Open your browser and go to
http://localhost:5000 - Click on the Elasticsearch button to download the data (will be downloaded in the
resourcesfolder) - Click on the OpenSearch button to index the data
- Click on the Server button to upload the agent
- Discuss with your agent in the chat window