03.1-semantic-search

Chapter 3.1 -- Semantic Search Hello World

Companion code for Generative AI in .NET, Chapter 3 sections 3.1.1 -- 3.1.5.

A six-row in-memory product catalog. Embeddings are generated once at startup (Ollama nomic-embed-text), then each query is embedded and ranked by cosine similarity. No vector store dependency -- intentionally minimal so the search math is visible.

Run it

ollama pull nomic-embed-text
dotnet run --project samples/ch03-rag/03.1-semantic-search

Try: wireless earbuds for running, cheap mouse, comfy mechanical keyboard.

Manuscript reference

• Manuscript/Chapter-03.md, sections 3.1.1 (text-to-vectors) and 3.1.5 (similarity search).
• Figure 3.1 ("From Text to Vectors") and 3.3 ("Semantic Search Query Flow").

Prerequisites

• .NET 9 SDK, Ollama with nomic-embed-text.