Skip to contents

create_rag_chain.R Overview

Source: R/create_rag_chain.R
create_rag_chain.Rd

A refined implementation of a LangChain-style Retrieval-Augmented Generation (RAG) pipeline. Includes vector search using DuckDB, optional web search using the Tavily API, and a built-in chat message history.

This function powers `create_rag_chain()`, the exported entry point for constructing a full RAG pipeline.

## Features: - Context-aware reformulation of user queries - Semantic chunk retrieval using DuckDB - Optional real-time web search (Tavily) - Compatible with any LLM function (OpenAI, Claude, etc.)

## Required Packages install.packages(c("DBI", "duckdb", "httr", "jsonlite", "stringi", "dplyr"))

Arguments

llm

A function that takes a prompt and returns a response (e.g. a call to OpenAI or Claude).

vector_database_directory

Path to the DuckDB database file.

method

Retrieval method backend. Currently only `"DuckDB"` is supported.

embedding_function

A function to embed text. Defaults to embed_openai().

system_prompt

Optional prompt with placeholders {chat_history}, {input}, {context}.

chat_history_prompt

Prompt used to rephrase follow-up questions using prior conversation history.

Tavily API key (set to NULL to disable web search).

embedding_dim

Integer; embedding vector dimension. Defaults to 1536.

Logical; whether to include web results from Tavily. Defaults to TRUE.

Value

A list of utility functions:

  • invoke(text) — Performs full context retrieval and LLM response

  • custom_invoke(text) — Retrieves context only (no LLM call)

  • get_session_history() — Returns complete conversation history

  • clear_history() — Clears in-memory chat history

  • disconnect() — Closes the underlying DuckDB connection

Details

Create a Retrieval-Augmented Generation (RAG) Chain

Creates a LangChain-style RAG chain using DuckDB for vector store operations, optional Tavily API for web search, and in-memory message history for conversational context.

Note

Only create_rag_chain() is exported. Helper functions are internal.

Examples

if (FALSE) { # \dontrun{
rag_chain <- create_rag_chain(
  llm = call_llm,
  vector_database_directory = "tests/testthat/test-data/my_vectors.duckdb",
  method = "DuckDB",
  embedding_function = embed_openai(),
  use_web_search = FALSE
)

response <- rag_chain$invoke("Tell me about R")
} # }