logion-client 0.1.11

Python client SDK for the Logion API

Logion Client SDK

Python SDK for the Logion API.

Installation

uv add logion-client

or

pip install logion-client

Quick Start

from logion import LogionClient

client = LogionClient(api_key="lgk_...")

# Health check
client.v1.health.check()

# Search course listings
client.v1.listings.search(query="rag")

# Get a specific course (UUID required)
client.v1.courses.get(course_id="550e8400-e29b-41d4-a716-446655440000")

# Create a user with an agent
client.v1.identity.create_user_with_agent(
    email="[e-mail]",
    user_password="secure-password",
    agent_name="My Agent",
)

# Top up credits (returns a Stripe Checkout URL)
client.v1.credits.create_top_up(amount_cents=1000)

# Purchase a course using credits
client.v1.courses.purchase(
    course_id="550e8400-e29b-41d4-a716-446655440000",
    expected_price_cents=500,
)

Configuration

The SDK reads LOGION_API_KEY and LOGION_BASE_URL from the environment by default. Explicit constructor arguments take precedence.

Parameter	Default	Environment variable	Description
api_key	"" (empty)	LOGION_API_KEY	API key for authentication
base_url	https://api.logion.sh	LOGION_BASE_URL	API base URL
timeout	30.0	—	Request timeout in seconds
max_retries	3	—	Max retry attempts (GET only)
extra_headers	{}	—	Additional headers per request

import os

# Using environment variables
os.environ["LOGION_API_KEY"] = "lgk_..."
client = LogionClient()

# Or passing explicitly
client = LogionClient(
    api_key="lgk_...",
    base_url="https://api.logion.sh",
    timeout=60.0,
    max_retries=5,
)

Error Handling

All errors inherit from LogionError. API errors provide structured details:

• LogionError — base exception for all SDK errors
• APIError — base for errors returned by the API (includes status_code, detail, request_id)
• AuthenticationError — 401 responses
• ForbiddenError — 403 responses
• NotFoundError — 404 responses
• ConflictError — 409 responses
• ValidationError — 422 responses
• RateLimitError — 429 responses
• ClientError — other 4xx responses not mapped above
• ServerError — 5xx responses
• TransportError — network failures (DNS, connection, timeout)

from logion import LogionClient, AuthenticationError, RateLimitError

client = LogionClient(api_key="lgk_...")

try:
    client.v1.courses.get(course_id="550e8400-e29b-41d4-a716-446655440000")
except AuthenticationError as exc:
    print(f"Auth failed: {exc.detail}")
except RateLimitError as exc:
    print(f"Rate limited: {exc.detail}")

Versioned Namespaces

The SDK organises endpoints by API version under the v1 namespace:

client.v1.health.check()
client.v1.listings.search(query="rag")

Future API versions will be accessible via client.v2, etc.

Mock Server Development

For local development and testing, you can point the SDK at a Prism mock server:

client = LogionClient(
    base_url="http://localhost:4010",
    api_key="lgk_test_mock_key",
)

Or set the LOGION_BASE_URL environment variable.

Code Generation

The SDK uses the OpenAPI spec as the source of truth for generated internals:

• Pydantic request and response models live under src/logion/v1/_types/generated/.
• Low-level HTTP operation functions live under src/logion/v1/_generated/.
• Public resource classes under src/logion/v1/_resources/ stay handwritten so the SDK remains ergonomic and stable.

# Generate models only
make generate-models

# Generate low-level operation functions only
make generate-operations

# Generate all client code derived from the OpenAPI contract
make generate-client

# Check if generated models are up to date
make check-models

# Check if generated operation functions are up to date
make check-operations

# Check all generated client code
make check-client

License

See the root repository LICENSE for details.