RAG-Enhanced Semantic Search

Uses LangGraph multi-agent workflow for Retrieval Augmented Generation. Provides question-answering over document corpus with source attribution.

Usage

run_rag_search(
  query,
  documents,
  ollama_model = "llama3",
  ollama_base_url = "http://localhost:11434",
  embedding_model = "nomic-embed-text",
  top_k = 5,
  envname = "textanalysisr-env"
)

Arguments

query

Character string, user question

documents

Character vector, corpus to search

ollama_model

Character string, LLM model (default: "llama3")

ollama_base_url

Character string, Ollama API endpoint

embedding_model

Character string, embedding model (default: "nomic-embed-text")

top_k

Integer, number of documents to retrieve (default: 5)

envname

Character string, Python environment name

Value

List with:

• success: Logical

• answer: Generated answer

• confidence: Confidence score (0-1)

• sources: Vector of source document IDs

• retrieved_docs: Retrieved document chunks

• scores: Similarity scores

Details

Multi-agent workflow:

1. Retrieval Agent: Find relevant documents via embeddings

2. Generation Agent: Create answer from context

3. Validation Agent: Assess answer quality

4. Conditional retry if confidence < 0.4

Requires Ollama with embedding model:

ollama pull llama3
ollama pull nomic-embed-text

Examples

if (FALSE) { # \dontrun{
documents <- c(
  "Assistive technology helps students with disabilities access curriculum.",
  "Universal Design for Learning provides multiple means of engagement.",
  "Response to Intervention uses tiered support systems."
)

result <- run_rag_search(
  query = "How does assistive technology support learning?",
  documents = documents
)

if (result$success) {
  cat("Answer:", result$answer, "\n")
  cat("Confidence:", result$confidence, "\n")
  cat("Sources:", paste(result$sources, collapse = ", "), "\n")
}
} # }