Skip to main content

A Gateway SDK for making API requests

Project description

Kosmoy Studio Logo

Kosmoy SDK

Python 3.7+ License: MIT PyPI

The Kosmoy SDK is a powerful Python library designed to seamlessly integrate your code with the Kosmoy platform. It empowers your applications to leverage Kosmoy Gateways, providing secure and governed access to AI models and routers configured within Kosmoy Studio.

This SDK is built with compatibility in mind, offering a similar interface to the standard OpenAI Python SDK and providing built-in integration with LangChain. This makes it easy to adopt the Kosmoy SDK, whether you are starting a new project or migrating an existing one.

Key Features

  • Secure Access to Kosmoy Gateways: Connect your Python applications to Kosmoy Gateways via a secure API key associated with a Kosmoy Studio Coded App.
  • Simplified Model Interaction: Interact with AI models (LLMs) registered in Kosmoy Studio using a familiar interface.
  • Leverage Studio Configurations: Benefit from the Routers and Guardrails configured within Kosmoy Studio, ensuring your application adheres to organizational policies.
  • OpenAI SDK Compatibility: Offers a similar interface to the standard OpenAI client for easy migration and adoption.
  • LangChain Integration: Provides seamless integration with LangChain for building sophisticated language model applications.
  • Comprehensive Error Handling: Implements robust error handling inherited from both the OpenAI and LangChain APIs.
  • Observability and Monitoring: Gain insights into your application's performance, usage, and costs through Kosmoy Studio's Insights.
  • Automatic Retries: Built-in retry mechanism for handling transient network issues.
  • Type Hints: Includes type hints for improved code readability and maintainability.

Installation

Standard Usage

pip install kosmoy-sdk

LangChain Integration

pip install kosmoy-sdk[langchain]

Requirements

  • Python 3.7+

Quick Start

Standard Usage

from kosmoy_sdk import GatewayClient


# Initialize the client
client = GatewayClient(
    app_id="your_app_id",    
    api_key="your_api_key"
    use_guardrails=True # Optionally enable guardrails [Message -> | Guardrail (inputs) | Model | Guardrail (output) | -> Response]
)

# Interact with a model (make sure the model is available in your Gateway)

response = client.client.chat.completions.create(
    model="gpt-4",  # Replace with your model name configured in Kosmoy Studio
    messages=[{"role": "user", "content": "Hello!"}],
)

# Get Gateway Information
gateway_info = client.get_gateway()

LangChain Integration

from kosmoy_sdk.langchain import KosmoyGatewayLangchain

# Initialize the LangChain client
client = KosmoyGatewayLangchain(
    app_id="your_app_id",  # Replace with your Coded App ID
    api_key="your_api_key",  # Replace with your API key
    model="gpt-4", # Replace with your model name configured in Kosmoy Studio,
    use_guardrails=True # Optionally enable guardrails [Message -> | Guardrail (inputs) | Model | Guardrail (output) | -> Response]
)

# Create messages
messages = [
    ("system", "You are a helpful assistant."),
    ("human", "What is the capital of France?")
]

# Get a response
response = client.invoke(messages)

# Get Gateway information
gateway_info = client.get_gateway()

Gateway information

gateway_info = client.get_gateway()
JSON Sample
{
  "id": 1,
  "name": "AI Text Generator",
  "description": "A coded app that generates text using AI models.",
  "gateway_id": 100,
  "created_at": "2025-02-10T12:00:00Z",
  "created_by_user": {
    "id": 5,
    "first_name": "John",
    "last_name": "Doe",
    "email": "john.doe@example.com",
    "role": "Admin"
  },
  "gateway": {
    "id": 100,
    "name": "gateway1",
    "description": "Gateway handling AI text processing requests.",
    "created_at": "2025-01-15T10:30:00Z",
    "created_by_user": {
      "id": 2,
      "first_name": "Alice",
      "last_name": "Smith",
      "email": "alice.smith@example.com",
      "role": "Root"
    },
    "models": [
      {
        "id": 201,
        "service_config_id": 301,
        "name": "gpt4o",
        "model_name": "gpt-4o",
        "description": "High-performance text generation model.",
        "config_params": {
          "temperature": 0.7,
          "max_tokens": 500
        }
      }
    ],
    "guardrails": [
      {
        "id": 10,
        "name": "guardrail1",
        "description": "Filters out inappropriate language."
      }
    ],
    "routers": [
      {
        "id": 50,
        "name": "routers1",
        "description": "Routes requests between different AI models.",
        "error_message": "Fallback to secondary model.",
        "primary_model_id": 201,
        "secondary_model_id": 202,
        "router_type": "FAILOVER",
        "created_by_user": {
          "id": 3,
          "first_name": "Bob",
          "last_name": "Johnson",
          "email": "bob.johnson@example.com",
          "role": "Admin"
        }
      }
    ]
  }
}

Configuration

The Kosmoy SDK requires the following parameters:

  • app_id: Your Kosmoy Studio Coded App ID (required).
  • api_key: Your Kosmoy Studio Coded App API key (required).
  • timeout: Request timeout in seconds (optional, default: 30).
  • max_retries: Maximum number of retry attempts (optional, default: 3).

How to get you Kosmoy Studio Coded App ID and API Key

  1. Log in to Kosmoy Studio: Access the Kosmoy Studio platform at https://www.kosmoy.com
  2. Create a Coded App: Navigate to the "Coded Apps" section and create a new Coded App.
  3. Assign a Gateway: Associate your Coded App with a configured Gateway.
  4. Retrieve Credentials: The API key and App ID will be displayed in the Coded App details.

Error Handling

The Kosmoy SDK inherits error handling from the OpenAI API (when used directly) and LangChain API (when using the LangChain integration). It also includes an automatic retry mechanism for failed requests.

Security

  • API credentials are transmitted securely via request headers.
  • The base URL for API communication is pre-configured within the SDK.
  • HTTPS is enforced for all API communications.

Learn More

License

This project is licensed under the MIT license.

Support

For support, please contact support@kosmoy.com

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

kosmoy_sdk-0.1.1.dev0.tar.gz (10.0 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

kosmoy_sdk-0.1.1.dev0-py3-none-any.whl (9.0 kB view details)

Uploaded Python 3

File details

Details for the file kosmoy_sdk-0.1.1.dev0.tar.gz.

File metadata

  • Download URL: kosmoy_sdk-0.1.1.dev0.tar.gz
  • Upload date:
  • Size: 10.0 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.1

File hashes

Hashes for kosmoy_sdk-0.1.1.dev0.tar.gz
Algorithm Hash digest
SHA256 7e024546f6362534338d337ccf189e1cf520fb641eafd71d0cb11376d772d59a
MD5 ae2a995f10deacf9f874b4a052de4dba
BLAKE2b-256 b0d9e80c1149f210240198da8e135aab2b395c97a5d480ce2882762fb0d0a0c9

See more details on using hashes here.

File details

Details for the file kosmoy_sdk-0.1.1.dev0-py3-none-any.whl.

File metadata

File hashes

Hashes for kosmoy_sdk-0.1.1.dev0-py3-none-any.whl
Algorithm Hash digest
SHA256 35f66eab1168c365cbed5650015f94ee7acc83663270a44653c20548ded04503
MD5 d6cd50eb068c4171885b9649a66431f9
BLAKE2b-256 a91f1292d045a0711707355d77dcf820c0a396d3d4767fc0a56e1a9dc3410f98

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page