infrastructure/vector_db_setup.md

## Vector Database Setup Guide

### 1. Faiss (Similarity Search)
Install: pip install faiss-cpu

```python
import faiss
import numpy as np

# Create index
index = faiss.IndexFlatL2(128)

# Add vectors
vectors = np.random.rand(100, 128).astype('float32')
index.add(vectors)
```

### 2. Weaviate
Install: docker run -d -p 8000:8000 weaviate/weaviate

Connect with:
```python
from weaviate import Weaviate
client = Weaviate.connect("http://localhost:8000")
client.create_class("Vectors")
```

### 3. Milvus
Install: pip install pymilvus

Start server: milvus-standalone-server

Connect:
```python
from pymilvus import connections, Collection
connections.connect("default", host="localhost", port="19530")
coll = Collection("vector_collection")
coll.insert([vectors])
```

### 4. Neo4j (Graph Storage)
Install: pip install neo4j

Connect:
```python
from neo4j import GraphDatabase
driver = GraphDatabase.driver("neo4j://localhost:7687", auth=basic_auth("neo4j", "password"))
with driver.session() as session:
    session.write_transaction(
        lambda tx: tx.run("CREATE (a:Vector {vector: $vector})", vector=vectors)
    )
```

Each system has unique features - choose based on your needs: Faiss for similarity search, Weaviate for semantic vectors, Milvus for large-scale storage, or Neo4j for graph-based relationships.