KISS AI Stack - Python Client SDK

The KISS AI Stack Client provides an easy-to-use interface for interacting with the KISS AI Stack Server stub, supporting RESTful and WebSocket APIs to manage Stack session lifecycle and execute tasks.

Features

  • REST client for session management, query execution, and document storage.
  • WebSocket client for real-time interactions with the Stack’s server session.

Getting Started

Nota Bene:
It is recommended to use this client in a proxy service for better security.

Installation

  1. Install the kiss-ai-stack-client package: 
    pip install kiss-ai-stack-client~=0.1.0a6
  2. Initialize the client:
  • REST API client
from kiss_ai_stack_client import RestEvent

# host name without mentioning protocol, i.e. example.com, localhost:8080
# secure_protocol, whether to use https or http
client = RestEvent(hostname="your-server-hostname", secure_protocol=True)
  • WebSocket-based client
from kiss_ai_stack_client import WebSocketEvent

# host name without mentioning protocol, i.e. example.com, localhost:8080
# secure_protocol, whether to use wss or ws
client = WebSocketEvent(hostname="your-server-hostname", secure_protocol=True)

Usage

1. Authorize a Stack Session

Create or refresh a Stack session:

To authorize the session, if you have saved a previous client_id and a client_secret, send them to get a new access token. Only persistent scope supports this. temporary scope will deactivate the client upon lifecycle ending.
Just send the scope only to generate a new client and keep the client_id and client_secret saved for persistent sessions.

session = await client.authorize_stack(scope="temporary")  # scopes - 'temporary', 'persistent'

# or, get a new access token for a previous 'persistent session'
session = await client.authorize_stack(client_id="your-client-id", client_secret="your-client-secret")

2. Bootstrap the Stack

Initialize the Stack session for task execution:

This step is a must to start the Stack session.

response = await client.bootstrap_stack(data="Hello, Stack!")

3. Generate an Answer

Send a query and receive the Stack’s response:

response = await client.generate_answer(data="What is the weather today?")

4. Store Documents

Upload files with optional metadata for storage:

files = ["path/to/document1.txt", "path/to/document2.pdf"]
metadata = {"category": "example"}
response = await client.store_data(files=files, metadata=metadata)
response = await client.generate_answer(data='Give me a summary of example documents')

5. Destroy the Stack Session

Close the current session (will clean up stored documents in a temporary session if it has):

response = await client.destroy_stack(data="Goodbye!")






KISS AI Stack. Copyright 2024.