how docs are added when using the agent_query method?

[n400peanuts]

I am trying to understand how documents are loaded, stored and embeddings created when using the agent_query method. This method works well and I would like to know what's working under the hood in order to save the embeddings created via this method.

[dosubot]

In the agent_query method, documents are handled as follows:

1. Loading Documents: If the docs parameter is not provided, a new Docs object is initialized. This object is used to manage the documents that will be processed.

2. Storing Documents: The method sets up a SearchIndex to store the answers and related data. This index is configured with fields such as "question" and uses a JSON model dump for storage. The index is named "answers" and is saved in the directory specified by query.settings.agent.index.index_directory.

3. Creating Embeddings: The embeddings are created and stored in the SearchIndex. The add_document method is used to add the document to the index, and the save_index method is used to save the index. This process involves embedding the query into a vector, ranking document chunks, and generating an answer using a language model ^[1].

This setup ensures that documents are properly loaded, stored, and that embeddings are created and saved efficiently.

To continue talking to Dosu, mention @dosu.