apiconfig 0.3.2.dev156

api config library

apiconfig

Flexible, extensible configuration and authentication for Python API clients.

---

Table of Contents

• Project Overview
• Quickstart
• Installation
• Key Features
• Usage
  • Basic Configuration
  • Authentication Strategies
  • Using Configuration Providers
  • Merging Configurations
• Practical Example: Real API Client Setup
• Logging
• Error Handling
• Testing and Coverage
• CI/CD
• Further Documentation
• License

---

Project Overview

apiconfig is a standalone Python library for managing API client configuration and authentication. It provides a robust, extensible foundation for building API clients, handling configuration (base URLs, timeouts, retries, headers) and supporting multiple authentication strategies (API key, Basic, Bearer, custom).

apiconfig is designed for:

• Developers building reusable, testable API clients.
• Projects needing flexible configuration sources (env, file, memory).
• Secure, pluggable authentication for HTTP APIs.

---

Quickstart

pip install apiconfig

from apiconfig import ClientConfig, ApiKeyAuth

auth = ApiKeyAuth(api_key="my-secret-key", header_name="X-API-Key")
config = ClientConfig(
    hostname="api.example.com",
    version="v1",
    auth_strategy=auth,
    timeout=10.0,
    retries=3,
)
print(config.base_url)  # https://api.example.com/v1

---

Installation

Install from PyPI:

pip install apiconfig

Or with Poetry:

poetry add apiconfig

---

Key Features

• Unified API Client Configuration: Manage base URLs, versions, headers, timeouts, retries, and more with a single, validated config object.
• Authentication Strategies: Built-in support for API Key, Basic, Bearer, and custom authentication via the Strategy Pattern.
• Config Providers: Load configuration from environment variables, files, or in-memory sources.
• Extensible: Easily add new authentication methods or config providers.
• Robust Error Handling: Clear, structured exception hierarchy for all config and auth errors.
• Type Safety: Strong type hints and type-checked public API.
• Logging Integration: Standard logging hooks for debugging and auditability.
• High Test Coverage: Around 94% coverage with unit and integration tests.

---

Usage

Basic Configuration

from apiconfig import ClientConfig

config = ClientConfig(
    hostname="api.example.com",
    version="v1",
    headers={"X-My-Header": "value"},
    timeout=10.0,
    retries=3,
)
print(config.base_url)  # https://api.example.com/v1

Authentication Strategies

API Key in Header

from apiconfig import ClientConfig, ApiKeyAuth

auth = ApiKeyAuth(api_key="my-secret-key", header_name="X-API-Key")
config = ClientConfig(
    hostname="api.example.com",
    version="v1",
    auth_strategy=auth,
)

API Key in Query Parameter

from apiconfig import ApiKeyAuth

auth = ApiKeyAuth(api_key="my-secret-key", param_name="api_key")

Basic and Bearer Auth

from apiconfig import BasicAuth, BearerAuth

basic = BasicAuth(username="user", password="pass")
bearer = BearerAuth(access_token="my-jwt-token")

Custom Authentication

from apiconfig import CustomAuth

custom = CustomAuth(auth_callable=lambda: {"Authorization": "Custom xyz"})

Using Configuration Providers

Environment Variables

from apiconfig import EnvProvider

env = EnvProvider(prefix="MYAPI_")
config_dict = env.load()
# Example: MYAPI_HOSTNAME=api.example.com, MYAPI_TIMEOUT=5

File and Memory Providers

from apiconfig import FileProvider, MemoryProvider

file_provider = FileProvider(file_path="config.json")
file_config = file_provider.load()

# MemoryProvider accepts configuration via the ``config_data`` parameter and
# exposes ``get_config`` instead of ``load``
memory_provider = MemoryProvider(config_data={"hostname": "api.example.com"})
memory_config = memory_provider.get_config()

Merging Configurations

from apiconfig import ClientConfig

base = ClientConfig(hostname="api.example.com", timeout=10)
override = ClientConfig(timeout=5, retries=2)
merged = base.merge(override)

---

Practical Example: Real API Client Setup

Below is a real-world example based on the integration tests. This pattern demonstrates how to use apiconfig to load configuration and secrets from environment variables, set up authentication, and make a request with an HTTP client (e.g., httpx):

import os
import httpx
from apiconfig import EnvProvider, ClientConfig, BearerAuth

# Load config and secrets from environment variables
env = EnvProvider()
config_dict = env.load()

# Get token and base URL from environment or use defaults
access_token = config_dict.get("FIKEN_ACCESS_TOKEN") or os.environ.get("FIKEN_ACCESS_TOKEN")
base_url = config_dict.get("FIKEN_BASE_URL") or os.environ.get("FIKEN_BASE_URL") or "https://api.fiken.no/api/v2"

# Set up authentication strategy if token is available
auth_strategy = BearerAuth(access_token) if access_token else None

# Create the API client configuration
client_config = ClientConfig(
    hostname=base_url,
    auth_strategy=auth_strategy,
)

# Prepare request headers using the auth strategy
headers = {}
if client_config.auth_strategy is not None:
    headers.update(client_config.auth_strategy.prepare_request_headers())

# Make a real HTTP request using httpx
with httpx.Client(timeout=10.0) as client:
    response = client.get(f"{client_config.base_url}/companies", headers=headers)
    print(response.status_code, response.json())

This approach can be adapted for any API and authentication method supported by apiconfig. See the integration tests for more real-world examples.

---

Logging

apiconfig uses standard Python logging. To enable debug output:

import logging

logging.basicConfig(level=logging.DEBUG)
logging.getLogger("apiconfig").setLevel(logging.INFO)

---

Error Handling

All errors are structured and documented. Common exceptions include:

• APIConfigError
• ConfigurationError
• AuthenticationError
• InvalidConfigError
• MissingConfigError
• AuthStrategyError

---

Testing and Coverage

apiconfig is fully tested with pytest and coverage.py. To run tests and check coverage:

pytest --cov=apiconfig --cov-report=html

---

CI/CD

Continuous integration and deployment are managed with GitHub Actions. All pushes and pull requests are tested, and releases are published to PyPI after passing tests.

---

Further Documentation

• Documentation (latest)
• CONTRIBUTING.md — Contribution guidelines
• Bug Tracker
• PyPI Project Page

---

License

LGPL-3.0-or-later. See LICENSE for details.