README
¶
Milvus — RAG with Milvus
Interactive RAG chat with Milvus as a vector store.
Note: Milvus is the most resource-intensive backend. Requires ~2 GB RAM to run. First startup may take time to initialize (start_period: 30s).
Quick start
1. Start Milvus (etcd + minio + milvus standalone):
docker compose up -d
2. Run the example:
cd examples/milvus && cp .env.example .env && go run .
For mock mode this is sufficient. For a real LLM, set LLM_PROVIDER=ollama|openai|anthropic and the corresponding keys.
Environment variables
| Variable | Default | Description |
|---|---|---|
LLM_PROVIDER |
mock |
LLM provider (mock, ollama, openai, anthropic) |
EMBEDDING_DIM |
1536 |
Vector dimension |
MILVUS_ADDR |
localhost:19121 |
Milvus REST API address |
COLLECTION_NAME |
draftrag_chunks |
Collection name |
Note
MilvusStore is an internal API (internal/infrastructure/vectorstore), status: "API in development".
The public API will be added in one of the upcoming releases.
Documentation
¶
Overview ¶
@sk-task docs-and-examples#T3.3: milvus example — RAG-чат с Milvus (AC-005). Использует публичный API draftrag напрямую. Shared только для mock/print. MilvusStore — внутренний API (internal/infrastructure/vectorstore), документирован как "API в разработке".
Быстрый старт с Docker:
docker compose up -d cp .env.example .env && go run .
Source Files