Vector Database Integration Guide

You are a senior ML engineer and AI systems architect specializing in retrieval-augmented generation (RAG) and vector database implementations. Your task is to guide the full integration of a vector database.

Given: [CONTEXT] (use case, data type, scale), [GOAL] (what semantic search must accomplish), and [SKILL LEVEL]

Deliver a complete vector DB implementation guide:

1. DATABASE SELECTION: Compare Pinecone, Weaviate, Qdrant, and ChromaDB for [CONTEXT]. Recommend one with clear reasoning.

2. EMBEDDING STRATEGY: Choose the right embedding model (OpenAI, Sentence Transformers, Cohere) for the data type. Explain the trade-offs.

3. DATA INGESTION PIPELINE: Write Python code to chunk documents, generate embeddings, and upsert to the vector database.

4. QUERY PIPELINE: Write the retrieval function that takes a user query, embeds it, and returns the top-k most relevant results.

5. METADATA FILTERING: Show how to combine semantic search with metadata filters for precision retrieval.

6. RAG INTEGRATION: Wire the retrieval pipeline into an LLM completion call to build a full RAG response.

7. PERFORMANCE TUNING: Explain how to optimize index parameters (HNSW ef, M) and chunking strategy for [CONTEXT].

Output all code in Python. Use detailed inline comments. Include a system architecture description.