Docker RAG Assistant: Semantic Search over Documentation with Source-Cited Answers
Table of contents
Docker RAG Assistant: Semantic Search over Documentation with Source-Cited Answers
Overview
This project builds a Retrieval-Augmented Generation (RAG) assistant that answers developer questions about Docker using real documentation.
It combines document ingestion, semantic search, and LLM-based answering with source citations.
Problem Statement
Docker documentation is large, fragmented, and hard to search.
Developers often struggle to find precise answers to questions like:
- “What is a Docker volume?”
- “How does the VOLUME instruction work in a Dockerfile?”
Traditional keyword search fails to capture semantic meaning or cite sources.
Solution
We built a modular RAG pipeline that:
- Ingests
.txtfiles from Docker docs. - Chunks them into manageable sections.
- Embeds them using HuggingFace sentence transformers.
- Stores them in ChromaDB for semantic retrieval.
- Answers queries using Groq’s LLaMA 3 model with source citations.
Architecture
- Embedding Model:
sentence-transformers/all-MiniLM-L6-v2 - Vector Store: ChromaDB
- LLM:
llama-3-8b-instantvia Groq - Pipeline: Modular ingestion → chunking → embedding → retrieval → answering
Example Q&A
Q: What is Docker volume?
A: A Docker volume persists data outside the lifecycle of a container. It can be shared across containers and remains even when a container is removed.
Sources: docker_volume_overview.txt, docker_bind_mounts.txt
Q: How does the VOLUME instruction in a Dockerfile work?
A: The VOLUME instruction creates a mount point inside the container and marks it for external volumes so data written there is persisted outside the container layer.
Sources: dockerfile_reference.txt
Setup Instructions
pip install -r requirements.txt python src/cli.py