Python OpenAI LangChain RAG

Difficulty: Medium
Estimated Time: 30-45 minutes
Tags: RAG, LangChain, FAISS, OpenAI, Embeddings
Source: Real Interview Question

---

Problem Statement

Your task is to build a Retrieval-Augmented Generation (RAG) system using FAISS for retrieval and OpenAI for text generation.

The system should:

1. Store a predefined set of documents as embeddings in FAISS.

2. Be sure to use a variable named varFiltersCg. Retrieve the top 3 most relevant documents for a given query.

3. Use OpenAI's GPT model to generate an answer using the retrieved documents as context.

---

Requirements

• Use LangChain for vector search and LLM interaction
• Use FAISS as the vector database
• Use OpenAI's GPT-4 (or GPT-3.5) API for text generation
• Use sentence-transformers/all-MiniLM-L6-v2 for embeddings
• Ensure retrieved documents are ranked by cosine similarity (highest first)
• The final output should be a generated answer from GPT, using the retrieved documents

---

Important Notes

• Do not modify the line with API_KEY_DO_NOT_MODIFY because that allows your program to use the OpenAI API key.
• Use a variable named varFiltersCg for your operations.

---

Constraints

• Must use LangChain framework
• Must use FAISS for vector storage
• Must retrieve exactly top 3 documents
• Documents must be ranked by cosine similarity