aice-data-brain

Python SDK for the Data Brain a multi-source connection layer that preprocesses, inspects, and answers questions over arbitrary data sources without exposing raw credentials to the model.

from aice_data_brain import DataBrainClient

client = DataBrainClient(base_url="https://db.example.com", api_key="sk-…")

# Connection lifecycle
client.connections.preprocess("postgres", config={"dsn": "…"})
client.connections.process("postgres", connection_id="cn_123")
client.connections.inspect("postgres", connection_id="cn_123")

# Query
result = client.connections.query(
    "postgres",
    connection_id="cn_123",
    query="SELECT * FROM events LIMIT 10",
)

# Describe what's connected
descriptors = client.descriptors.list()

Endpoints covered (see data_brain/app/api/routes/):

• GET /health
• POST /connections/{connection_type}/preprocess
• POST /connections/{connection_type}/process
• POST /connections/{connection_type}/cleanup
• POST /connections/{connection_type}/query
• POST /connections/{connection_type}/inspect
• GET /descriptors/… (list available data descriptors)