This shows you the differences between two versions of the page.
| Next revision | Previous revision | ||
| wiki:ai:rag-app [2025/08/05 19:19] – created mcarver | wiki:ai:rag-app [2025/08/06 14:06] (current) – [How It Works] mcarver | ||
|---|---|---|---|
| Line 38: | Line 38: | ||
| ---- | ---- | ||
| - | ===== How It Works ===== | ||
| - | |||
| - | ==== 1. Document Ingestion ==== | ||
| - | - Connects to your Azure Blob Storage | ||
| - | - Parses `.txt`, `.docx`, and other file types | ||
| - | - Splits documents into manageable chunks | ||
| - | - Generates embeddings using OpenAI (e.g., `text-embedding-ada-002`) | ||
| - | - Indexes both text and vectors into Azure Cognitive Search | ||
| - | |||
| - | ==== 2. Query Workflow ==== | ||
| - | - User submits a question | ||
| - | - App: | ||
| - | - Generates embedding for the question | ||
| - | - Searches Cognitive Search for top-K relevant chunks | ||
| - | - Constructs a prompt with context + question | ||
| - | - Uses GPT model to answer the question | ||
| - | - Returns answer along with source documents | ||
| - | |||
| - | ---- | ||
| ===== Use Cases ===== | ===== Use Cases ===== | ||