RAG Knowledge Base

# retriever.py - the full RAG pipeline in one function
  def ask(question: str) -> str:
      # 1. Embed the question in the same vector space as the chunks
      docs = vectorstore.similarity_search(question, k=TOP_K)
  
      if not docs:
          return "No relevant documents found."
  
      # 2. Build context from top-4 retrieved chunks
      context = "\n\n".join(d.page_content for d in docs)
  
      # 3. GPT-4o-mini answers ONLY from context - not from training data
      response = openai.chat.completions.create(
          model="gpt-4o-mini",
          messages=[
              {"role": "system", "content":
                  "Answer only using the context below. "
                  "If the answer is not in the context, say so."},
              {"role": "user", "content":
                  f"Context:\n{context}\n\nQuestion: {question}"}
          ]
      )
  
      # 4. Deduplicate sources and append to answer
      sources = list({
          f"{d.metadata['source']} (page {d.metadata['page']})"
          for d in docs
      })
      return response.choices[0].message.content + "\n\n📎 Sources:\n" + "\n".join(f" • {s}" for s in sources)