Cross-platform layered configuration loader for Python

These details have not been verified by PyPI

Project links

Project description

lib_layered_config

A cross-platform configuration loader that deep-merges application defaults, host overrides, user profiles, .env files, and environment variables into a single immutable object. The core follows Clean Architecture boundaries so adapters (filesystem, dotenv, environment) stay isolated from the domain model while the CLI mirrors the same orchestration.

Key Features
Architecture Overview
Installation
Quick Start
Understanding Key Identifiers: Vendor, App, and Slug
Configuration File Structure
Configuration Sources & Precedence
CLI Usage
Python API
Example Generation & Deployment
Provenance & Observability
Development
License

Key Features

Deterministic layering — precedence is always defaults → app → host → user → dotenv → env.
Immutable value object — returned Config prevents accidental mutation and exposes dotted-path helpers.
Provenance tracking — every key reports the layer and path that produced it.
Cross-platform path discovery — Linux (XDG), macOS, and Windows layouts with environment overrides for tests.
Extensible formats — TOML and JSON are built-in; YAML is available via the optional yaml extra.
Automation-friendly CLI — inspect, deploy, or scaffold configurations without writing Python.
Structured logging — adapters emit trace-aware events without polluting the domain layer.

Architecture Overview

The project follows a Clean Architecture layout so responsibilities remain easy to reason about and test:

Domain — immutable Config value object plus error taxonomy.
Application — merge policy (LayerSnapshot, merge_layers) and adapter protocols.
Adapters — filesystem discovery, structured file loaders, dotenv, and environment ingress.
Composition — core and _layers wire adapters together and expose the public API.
Presentation & Tooling — CLI commands, deployment/example helpers, observability utilities, and testing hooks.

Consult docs/systemdesign/module_reference.md for a per-module catalogue and traceability back to the system design notes.

Installation

pip install lib_layered_config
# or with optional YAML support
pip install "lib_layered_config[yaml]"

Requires Python 3.13+ — the standard-library tomllib handles TOML parsing.

Install the optional yaml extra only when you actually ship .yml files to keep the dependency footprint small.

For local development add tooling extras:

pip install "lib_layered_config[dev]"

Quick Start

from lib_layered_config import read_config

config = read_config(vendor="Acme", app="ConfigKit", slug="config-kit")
print(config.get("service.timeout", default=30))
print(config.origin("service.timeout"))

CLI equivalent (human readable by default):

lib_layered_config read --vendor Acme --app ConfigKit --slug config-kit

JSON output including provenance:

lib_layered_config read --vendor Acme --app ConfigKit --slug config-kit --format json
# or
lib_layered_config read-json --vendor Acme --app ConfigKit --slug config-kit

Understanding Key Identifiers: Vendor, App, and Slug

Before diving into configuration sources, it's important to understand the three key identifiers used throughout this library:

Vendor

What it is: Your organization or company name (e.g., "Acme", "Mozilla", "MyCompany").

Where it's used:

macOS: /Library/Application Support/Acme/MyApp/ and ~/Library/Application Support/Acme/MyApp/
Windows: C:\ProgramData\Acme\MyApp\ and %APPDATA%\Acme\MyApp\
Linux: Not used (Linux uses the slug directly)

Example:

# Your company is "Acme Corp"
config = read_config(vendor="Acme", app="DatabaseTool", slug="db-tool")
# macOS paths: /Library/Application Support/Acme/DatabaseTool/config.toml

App

What it is: Your application's full/display name (e.g., "DatabaseTool", "ConfigKit", "MyService").

Where it's used:

macOS: Combined with vendor in paths: /Library/Application Support/Acme/MyApp/
Windows: Combined with vendor: C:\ProgramData\Acme\MyApp\
Linux: Not used (Linux uses the slug directly)
Default slug: If you don't specify a slug, the app name is used as the slug

Example:

config = read_config(vendor="Acme", app="ConfigKit", slug="config-kit")
# macOS: /Library/Application Support/Acme/ConfigKit/config.toml
# Windows: C:\ProgramData\Acme\ConfigKit\config.toml

Slug (Configuration Slug)

What it is: A lowercase, filesystem-friendly identifier for your configuration (e.g., "myapp", "config-kit", "db-tool").

Why it exists: The slug serves as a universal, platform-independent identifier for your configuration that works consistently across:

Linux/UNIX filesystem paths (case-sensitive, prefers hyphens)
Environment variable prefixes (converted to uppercase)
Cross-platform scripts and automation

Where it's used:

1. Linux/UNIX Paths

/etc/xdg/myapp/config.toml               # System-wide (XDG-compliant)
/etc/xdg/myapp/hosts/server-01.toml      # Host-specific (XDG-compliant)
~/.config/myapp/config.toml              # User-specific
~/.config/myapp/.env                      # Environment variables

Note: For backwards compatibility, the library also checks /etc/myapp/ if /etc/xdg/myapp/ is not found.

2. Environment Variable Prefix

The slug is converted to uppercase with underscores, followed by a triple underscore (___) separator to clearly distinguish the prefix from section/key separators (which use double underscores __):

# Slug: "myapp" → Environment prefix: "MYAPP___"
MYAPP___DATABASE__HOST=localhost
MYAPP___DATABASE__PORT=5432
MYAPP___SERVICE__TIMEOUT=30

# Slug: "config-kit" → Environment prefix: "CONFIG_KIT___"
CONFIG_KIT___API__KEY=secret
CONFIG_KIT___DEBUG__ENABLED=true

3. Cross-Platform Consistency

The slug provides a consistent identifier regardless of platform:

# Same slug works on all platforms
config = read_config(vendor="Acme", app="My App", slug="myapp")

# Linux:   /etc/xdg/myapp/config.toml
# macOS:   /Library/Application Support/Acme/My App/config.toml
# Windows: C:\ProgramData\Acme\My App\config.toml
# Env vars: MYAPP___DATABASE__HOST (all platforms)

Slug Naming Best Practices

✅ DO:

Use lowercase letters: "myapp", "database-tool"
Use hyphens for word separation: "config-kit", "db-manager"
Keep it short and memorable: "myapp" not "my-super-awesome-application"
Use the same slug everywhere in your application

❌ DON'T:

Use spaces: "my app" → use "myapp" or "my-app"
Use uppercase: "MyApp" → use "myapp"
Use underscores in the slug: "my_app" → use "my-app" (underscores are added automatically for env vars)
Mix naming conventions across your codebase
Use path separators (/ or \): "../etc" will raise ValueError
Start with a dot: ".hidden" will raise ValueError

Complete Example: How They Work Together

from lib_layered_config import read_config

# Define your application identity
config = read_config(
    vendor="Acme",           # Your company name
    app="DatabaseManager",   # Your application's display name
    slug="db-manager"        # Filesystem/environment-friendly identifier
)

This creates the following structure:

On Linux:

/etc/xdg/db-manager/config.toml          # System-wide (uses slug, XDG-compliant)
~/.config/db-manager/config.toml         # User-specific (uses slug)
Environment: DB_MANAGER___*              # Env prefix (slug → uppercase + ___)

On macOS:

/Library/Application Support/Acme/DatabaseManager/config.toml   # System-wide (vendor + app)
~/Library/Application Support/Acme/DatabaseManager/config.toml  # User-specific (vendor + app)
Environment: DB_MANAGER___*                                      # Env prefix (slug → uppercase + ___)

On Windows:

C:\ProgramData\Acme\DatabaseManager\config.toml     # System-wide (vendor + app)
%APPDATA%\Acme\DatabaseManager\config.toml          # User-specific (vendor + app)
Environment: DB_MANAGER___*                         # Env prefix (slug → uppercase + ___)

Why Three Identifiers?

Different platforms have different conventions:

Windows/macOS: Prefer human-readable names with spaces and mixed case ("Acme Corp", "My Application")
Linux/UNIX: Prefer lowercase with hyphens (myapp, config-kit)
Environment variables: Must use uppercase with underscores (MYAPP_, CONFIG_KIT_)

This library uses three identifiers so your application can follow native conventions on each platform while maintaining a consistent configuration identity.

Quick Reference Table

Identifier	Format	Example	Used In
vendor	Mixed case, spaces OK	`"Acme Corp"`	macOS, Windows paths
app	Mixed case, spaces OK	`"Database Manager"`	macOS, Windows paths
slug	lowercase-with-hyphens	`"db-manager"`	Linux paths, env var prefix (becomes `DB_MANAGER___`)

Configuration File Structure

Configuration files use TOML, JSON, or YAML format. Here's a comprehensive example showing how to structure your configuration:

Example Configuration File (TOML)

# config.toml - Complete example showing all structural features

# ============================================================================
# TOP-LEVEL KEYS (Simple Values)
# ============================================================================
# Access in Python: config.get("debug")
# Access in CLI: config["debug"]
# Environment variable: MYAPP___DEBUG=true

debug = false
environment = "production"
version = "1.0.0"


# ============================================================================
# SECTIONS (Tables in TOML)
# ============================================================================
# Sections group related configuration values
# Access in Python: config.get("database.host")
# Environment variable: MYAPP___DATABASE__HOST=localhost

[database]
host = "localhost"
port = 5432
name = "myapp_db"
username = "admin"
# Passwords should come from environment variables or .env files!
# Set via: MYAPP___DATABASE__PASSWORD=secret


# ============================================================================
# NESTED SECTIONS (Subtables)
# ============================================================================
# Use dot notation for nested sections
# Access in Python: config.get("database.pool.size")
# Environment variable: MYAPP___DATABASE__POOL__SIZE=20

[database.pool]
size = 10
max_overflow = 20
timeout = 30
recycle = 3600  # seconds


[database.ssl]
enabled = true
verify = true
cert_path = "/etc/ssl/certs/db.pem"


# ============================================================================
# ARRAYS (Lists)
# ============================================================================
# Access in Python: config.get("database.replicas")
# Returns: ["replica1.db.local", "replica2.db.local"]

[database]
replicas = [
    "replica1.db.local",
    "replica2.db.local",
    "replica3.db.local"
]


# ============================================================================
# SERVICE CONFIGURATION
# ============================================================================

[service]
name = "MyApp Service"
host = "0.0.0.0"
port = 8080
timeout = 30
base_url = "https://api.example.com"


# Nested retry configuration
[service.retry]
max_attempts = 3
backoff_multiplier = 2
initial_delay = 1.0  # seconds


# Multiple endpoints
[service.endpoints]
api = "/api/v1"
health = "/health"
metrics = "/metrics"


# ============================================================================
# LOGGING CONFIGURATION
# ============================================================================

[logging]
level = "INFO"  # DEBUG, INFO, WARNING, ERROR, CRITICAL
format = "json"
output = "stdout"


[logging.handlers]
console = true
file = true
syslog = false


[logging.files]
path = "/var/log/myapp/app.log"
max_bytes = 10485760  # 10 MB
backup_count = 5


# ============================================================================
# FEATURE FLAGS
# ============================================================================
# Use sections for grouping related feature flags

[features]
new_ui = false
experimental_api = false
beta_features = false


[features.analytics]
enabled = true
sampling_rate = 0.1  # 10% of requests


# ============================================================================
# API CONFIGURATION
# ============================================================================

[api]
rate_limit = 1000  # requests per minute
timeout = 60


[api.authentication]
type = "jwt"  # jwt, oauth, apikey
token_expiry = 3600  # seconds


[api.cors]
enabled = true
allowed_origins = ["https://example.com", "https://app.example.com"]
allowed_methods = ["GET", "POST", "PUT", "DELETE"]


# ============================================================================
# CACHE CONFIGURATION
# ============================================================================

[cache]
backend = "redis"  # redis, memcached, memory
default_ttl = 300  # seconds


[cache.redis]
host = "localhost"
port = 6379
db = 0
password = ""  # Set via MYAPP___CACHE__REDIS__PASSWORD


# ============================================================================
# EMAIL CONFIGURATION
# ============================================================================

[email]
enabled = true
from_address = "noreply@example.com"
from_name = "MyApp"


[email.smtp]
host = "smtp.gmail.com"
port = 587
use_tls = true
username = "notifications@example.com"
# Password should come from environment: MYAPP___EMAIL__SMTP__PASSWORD


# ============================================================================
# MONITORING & OBSERVABILITY
# ============================================================================

[monitoring]
enabled = true


[monitoring.metrics]
backend = "prometheus"
port = 9090
path = "/metrics"


[monitoring.tracing]
enabled = true
backend = "jaeger"
sample_rate = 0.01  # 1% of requests


[monitoring.healthcheck]
enabled = true
interval = 30  # seconds

Understanding the Structure

1. Top-Level Keys

Simple key-value pairs at the root level:

debug = false
environment = "production"

Access:

Python: config.get("debug") or config["debug"]
CLI: Value appears as debug: false
Environment: MYAPP___DEBUG=true

2. Sections (Tables)

Sections group related configuration:

[database]
host = "localhost"
port = 5432

Access:

Python: config.get("database.host") → "localhost"
Python: config["database"] → {"host": "localhost", "port": 5432}
Environment: MYAPP___DATABASE__HOST=postgres.local

3. Nested Sections (Subtables)

Use dot notation for deeper nesting:

[database.pool]
size = 10
timeout = 30

[database.ssl]
enabled = true
verify = true

Access:

Python: config.get("database.pool.size") → 10
Python: config.get("database.ssl.enabled") → true
Environment: MYAPP___DATABASE__POOL__SIZE=20
Environment: MYAPP___DATABASE__SSL__ENABLED=false

Structure visualization:

{
    "database": {
        "host": "localhost",
        "port": 5432,
        "pool": {
            "size": 10,
            "timeout": 30
        },
        "ssl": {
            "enabled": true,
            "verify": true
        }
    }
}

4. Arrays (Lists)

Multiple values in a list:

[database]
replicas = [
    "replica1.db.local",
    "replica2.db.local"
]

[api.cors]
allowed_origins = ["https://example.com", "https://app.example.com"]

Access:

Python: config.get("database.replicas") → ["replica1.db.local", "replica2.db.local"]
Python: config.get("database.replicas")[0] → "replica1.db.local"

5. Data Types

TOML supports multiple data types:

# Strings
name = "MyApp"
host = "localhost"

# Integers
port = 8080
timeout = 30

# Floats
sample_rate = 0.1
backoff_multiplier = 2.5

# Booleans
debug = false
enabled = true

# Arrays
replicas = ["host1", "host2"]
allowed_methods = ["GET", "POST"]

# Dates/Times (TOML feature)
created_at = 2024-01-15T10:30:00Z

Complete Python Access Example

Given the example configuration above, here's how to access values:

from lib_layered_config import read_config

config = read_config(vendor="Acme", app="MyApp", slug="myapp")

# Top-level keys
debug = config.get("debug")  # false
env = config.get("environment")  # "production"

# Section values
db_host = config.get("database.host")  # "localhost"
db_port = config.get("database.port")  # 5432

# Nested section values
pool_size = config.get("database.pool.size")  # 10
ssl_enabled = config.get("database.ssl.enabled")  # true

# Deep nesting
retry_attempts = config.get("service.retry.max_attempts")  # 3
smtp_port = config.get("email.smtp.port")  # 587

# Arrays
replicas = config.get("database.replicas")  # ["replica1.db.local", ...]
first_replica = config.get("database.replicas")[0]  # "replica1.db.local"

# Feature flags
new_ui_enabled = config.get("features.new_ui")  # false
analytics_rate = config.get("features.analytics.sampling_rate")  # 0.1

# With defaults
api_timeout = config.get("api.timeout", default=60)  # 60
cache_ttl = config.get("cache.default_ttl", default=300)  # 300

Environment Variable Mapping

The slug is converted to an environment prefix (uppercase with triple underscore ___ separator), and nested keys use double underscores (__):

# Slug: "myapp" → Prefix: "MYAPP___"

# Top-level keys
MYAPP___DEBUG=true
MYAPP___ENVIRONMENT=staging

# Section keys (triple underscore after prefix, double for nesting)
MYAPP___DATABASE__HOST=postgres.production.local
MYAPP___DATABASE__PORT=5433

# Nested sections (each level separated by __)
MYAPP___DATABASE__POOL__SIZE=50
MYAPP___DATABASE__SSL__ENABLED=true

# Deep nesting
MYAPP___SERVICE__RETRY__MAX_ATTEMPTS=5
MYAPP___EMAIL__SMTP__PASSWORD=secret123
MYAPP___FEATURES__ANALYTICS__SAMPLING_RATE=0.5

# Arrays (JSON format for complex types)
MYAPP___DATABASE__REPLICAS='["replica1", "replica2"]'

Key Pattern:

Prefix: SLUG in uppercase followed by ___
Separator: Triple underscore (___) after prefix to distinguish from nesting
Nesting: Double underscores (__) for each level
Format: PREFIX___SECTION__SUBSECTION__KEY=value

JSON and YAML Equivalents

The same structure in JSON:

{
  "debug": false,
  "environment": "production",
  "database": {
    "host": "localhost",
    "port": 5432,
    "pool": {
      "size": 10,
      "timeout": 30
    },
    "ssl": {
      "enabled": true,
      "verify": true
    },
    "replicas": ["replica1.db.local", "replica2.db.local"]
  },
  "service": {
    "host": "0.0.0.0",
    "port": 8080,
    "retry": {
      "max_attempts": 3,
      "backoff_multiplier": 2
    }
  }
}

And in YAML:

debug: false
environment: production

database:
  host: localhost
  port: 5432
  pool:
    size: 10
    timeout: 30
  ssl:
    enabled: true
    verify: true
  replicas:
    - replica1.db.local
    - replica2.db.local

service:
  host: 0.0.0.0
  port: 8080
  retry:
    max_attempts: 3
    backoff_multiplier: 2

All three formats produce the same configuration structure and can be accessed identically through the library.

Configuration Sources & Precedence

Later layers override earlier ones per key while leaving unrelated keys untouched.

Precedence	Layer	Description
0	`defaults`	Optional baseline file provided via the API/CLI `--default-file` flag
1	`app`	System-wide defaults (e.g. `/etc/<slug>/…`)
2	`host`	Machine-specific overrides (`hosts/<hostname>.toml`)
3	`user`	Per-user settings (XDG, Application Support, AppData)
4	`dotenv`	First `.env` found via upward search plus platform extras
5	`env`	Process environment with namespacing and `__` nesting

Use the optional defaults layer when you want one explicitly-provided file to seed configuration before host/user overrides apply.

Important directories (overridable via environment variables):

Linux

/etc/<slug>/config.toml
/etc/<slug>/config.d/*.{toml,json,yaml,yml}
/etc/<slug>/hosts/<hostname>.toml
$XDG_CONFIG_HOME/<slug>/config.toml (the resolver reads $XDG_CONFIG_HOME; if it is missing it falls back to ~/.config/<slug>/config.toml)
.env search: current directory upwards + $XDG_CONFIG_HOME/<slug>/.env

macOS

/Library/Application Support/<Vendor>/<App>/config.toml
/Library/Application Support/<Vendor>/<App>/config.d/
/Library/Application Support/<Vendor>/<App>/hosts/<hostname>.toml
~/Library/Application Support/<Vendor>/<App>/config.toml
.env search: current directory upwards + ~/Library/Application Support/<Vendor>/<App>/.env

Windows

%ProgramData%\<Vendor>\<App>\config.toml
%ProgramData%\<Vendor>\<App>\config.d\*
%ProgramData%\<Vendor>\<App>\hosts\%COMPUTERNAME%.toml
%APPDATA%\<Vendor>\<App>\config.toml (resolver order: LIB_LAYERED_CONFIG_APPDATA → %APPDATA%; if neither yields an existing directory it tries LIB_LAYERED_CONFIG_LOCALAPPDATA → %LOCALAPPDATA%)
.env search: current directory upwards + %APPDATA%\<Vendor>\<App>\.env

Environment overrides: LIB_LAYERED_CONFIG_ETC, LIB_LAYERED_CONFIG_PROGRAMDATA, LIB_LAYERED_CONFIG_APPDATA, LIB_LAYERED_CONFIG_LOCALAPPDATA, LIB_LAYERED_CONFIG_MAC_APP_ROOT, LIB_LAYERED_CONFIG_MAC_HOME_ROOT. Both the runtime readers and the deploy helper honour these variables so generated files land in the same directories that read_config inspects.

Fallback note: Whenever a path is marked as a fallback, the resolver first consults the documented environment overrides (LIB_LAYERED_CONFIG_*, $XDG_CONFIG_HOME, %APPDATA%, etc.). If those variables are unset or the computed directory does not exist, it switches to the stated fallback location (~/.config, %LOCALAPPDATA%, ...). This keeps local installs working without additional environment configuration while still allowing operators to steer resolution explicitly.

CLI Usage

Command Summary

Command	Description
`lib_layered_config read`	Load configuration (human readable by default)
`lib_layered_config read-json`	Emit config + provenance JSON envelope
`lib_layered_config deploy`	Copy a source file into one or more layer directories
`lib_layered_config generate-examples`	Scaffold example trees (POSIX/Windows layouts)
`lib_layered_config env-prefix`	Compute the canonical environment prefix
`lib_layered_config info`	Print package metadata
`lib_layered_config fail`	Intentionally raise a `RuntimeError` (for testing)

`read`

Load configuration and print either human-readable prose or JSON.

Usage:

lib_layered_config read --vendor Acme --app ConfigKit --slug config-kit \
  [--prefer toml] [--prefer json] \
  [--start-dir /path/to/project] \
  [--default-file ./config.defaults.toml] \
  [--format human|json] \
  [--indent | --no-indent] \
  [--provenance | --no-provenance]

Parameters:

Parameter	Type	Required	Default	Description
`--vendor`	string	Yes	-	Vendor namespace used to compute filesystem paths
`--app`	string	Yes	-	Application name used to compute filesystem paths
`--slug`	string	Yes	-	Configuration slug for file paths and environment prefix
`--prefer`	string	No	None	Preferred file suffix (repeatable flag: `--prefer toml --prefer json`). Earlier values take precedence. Valid values: `toml`, `json`, `yaml`, `yml`
`--start-dir`	path	No	current dir	Starting directory for upward `.env` file search. Must be an existing directory
`--default-file`	path	No	None	Path to lowest-precedence defaults file. Must be an existing file
`--format`	choice	No	`human`	Output format. Valid values: `human` (annotated prose), `json` (structured JSON)
`--indent` / `--no-indent`	flag	No	`--indent`	Pretty-print JSON output with indentation. Only applies when `--format json`
`--provenance` / `--no-provenance`	flag	No	`--provenance`	Include provenance metadata in JSON output. Only applies when `--format json`

Examples:

Example 1: Basic configuration inspection (human-readable)

# Load and display configuration in human-readable format
lib_layered_config read --vendor Acme --app MyApp --slug myapp

Output:

service.timeout: 30
  provenance: layer=app, path=/etc/xdg/myapp/config.toml
service.endpoint: https://api.example.com
  provenance: layer=user, path=/home/alice/.config/myapp/config.toml
database.host: localhost
  provenance: layer=env, path=None
database.port: 5432
  provenance: layer=app, path=/etc/xdg/myapp/config.toml

Explanation: The default format shows each configuration value with its source layer and file path (or "None" for environment variables). Perfect for quick debugging.

Example 2: JSON output for automation scripts

# Get configuration as JSON for use in shell scripts
config_json=$(lib_layered_config read \
  --vendor Acme --app MyApp --slug myapp \
  --format json --no-provenance --no-indent)

# Parse with jq
echo "$config_json" | jq -r '.database.host'
# Output: localhost

Explanation: Use --format json --no-provenance --no-indent to get just the configuration values as compact JSON, perfect for piping to jq or other JSON processors.

Example 3: Full audit with provenance (JSON)

# Get both configuration and provenance metadata
lib_layered_config read \
  --vendor Acme --app MyApp --slug myapp \
  --format json --provenance --indent > config-audit.json

# View the structure
cat config-audit.json

Output:

{
  "config": {
    "service": {
      "timeout": 30,
      "endpoint": "https://api.example.com"
    },
    "database": {
      "host": "localhost",
      "port": 5432
    }
  },
  "provenance": {
    "service.timeout": {
      "layer": "app",
      "path": "/etc/xdg/myapp/config.toml",
      "key": "service.timeout"
    },
    "service.endpoint": {
      "layer": "user",
      "path": "/home/alice/.config/myapp/config.toml",
      "key": "service.endpoint"
    },
    "database.host": {
      "layer": "env",
      "path": null,
      "key": "database.host"
    }
  }
}

Explanation: This gives you complete audit information - both the final configuration values and where each one came from.

Example 4: Using file format preferences

# Prefer TOML files, then JSON, then YAML
lib_layered_config read \
  --vendor Acme --app MyApp --slug myapp \
  --prefer toml --prefer json --prefer yaml

Explanation: When multiple configuration file formats exist in the same directory (e.g., config.toml and config.json), the --prefer flag controls which one takes precedence. Earlier values win.

Example 5: Load with defaults and specific .env location

# Load configuration with shipped defaults and project-specific .env
lib_layered_config read \
  --vendor Acme --app MyApp --slug myapp \
  --default-file ./config/defaults.toml \
  --start-dir /opt/myapp \
  --format human

Explanation: Use --default-file to provide application defaults that ship with your app, and --start-dir to specify where to start searching for .env files (useful when running from a different directory).

Example 6: Debugging configuration issues

# Check if environment variables are overriding your config
MYAPP___SERVICE__TIMEOUT=5 lib_layered_config read \
  --vendor Acme --app MyApp --slug myapp \
  --format human | grep -A1 "service.timeout"

Output:

service.timeout: 5
  provenance: layer=env, path=None

Explanation: Set environment variables before the command to test how they override file-based configuration. The provenance shows which layer won.

`read-json`

Always emit combined JSON output (config + provenance). This is a convenience alias for read --format json --provenance.

Usage:

lib_layered_config read-json --vendor Acme --app ConfigKit --slug config-kit \
  [--prefer toml] [--prefer json] \
  [--start-dir /path/to/project] \
  [--default-file ./config.defaults.toml] \
  [--indent | --no-indent]

Parameters:

Parameter	Type	Required	Default	Description
`--vendor`	string	Yes	-	Vendor namespace
`--app`	string	Yes	-	Application name
`--slug`	string	Yes	-	Configuration slug
`--prefer`	string	No	None	Preferred file suffix (repeatable)
`--start-dir`	path	No	current dir	Starting directory for `.env` search
`--default-file`	path	No	None	Path to defaults file
`--indent` / `--no-indent`	flag	No	`--indent`	Pretty-print JSON output

Example:

lib_layered_config read-json --vendor Acme --app ConfigKit --slug config-kit --indent

`deploy`

Copy a source configuration file into one or more layer directories.

Usage:

lib_layered_config deploy --source ./config/app.toml \
  --vendor Acme --app ConfigKit --slug config-kit \
  --target app [--target host] [--target user] \
  [--platform linux|darwin|windows] \
  [--force | --no-force]

Parameters:

Parameter	Type	Required	Default	Description
`--source`	path	Yes	-	Path to the configuration file to copy. Must be an existing file
`--vendor`	string	Yes	-	Vendor namespace
`--app`	string	Yes	-	Application name
`--slug`	string	Yes	-	Configuration slug
`--target`	choice	Yes	-	Layer targets to deploy to (repeatable flag). Valid values: `app`, `host`, `user`. Can specify multiple: `--target app --target user`
`--platform`	string	No	auto-detect	Override platform. Valid values: `linux`, `darwin`, `windows`, or any string starting with `win`
`--force` / `--no-force`	flag	No	`--no-force`	Overwrite existing files at destinations

Returns: JSON array of file paths created or overwritten.

🔒 File Overwrite Behavior

The deploy command has safe-by-default behavior to prevent accidental data loss:

Default Behavior (without `--force`):

✅ Creates new files if they don't exist
❌ Skips existing files - will NOT overwrite
📋 Returns empty array [] or partial array if some files were skipped
🛡️ Protects user customizations from being accidentally overwritten

# First deployment - creates file
lib_layered_config deploy --source ./config.toml \
  --vendor Acme --app MyApp --slug myapp --target user
# Output: ["/home/alice/.config/myapp/config.toml"]

# Second deployment (same command) - skips existing file
lib_layered_config deploy --source ./config.toml \
  --vendor Acme --app MyApp --slug myapp --target user
# Output: []  ← File already exists, not overwritten

With `--force` Flag:

✅ Creates new files if they don't exist
✅ Overwrites existing files without warning
📋 Returns array of all files created/overwritten
⚠️ Use with caution - existing content will be lost

# Force overwrite existing files
lib_layered_config deploy --source ./config.toml \
  --vendor Acme --app MyApp --slug myapp --target user --force
# Output: ["/home/alice/.config/myapp/config.toml"]  ← Overwritten

Decision Flow Diagram

┌─────────────────────────────────┐
│  lib_layered_config deploy      │
│  --source config.toml            │
│  --target user                   │
└────────────┬────────────────────┘
             │
             ▼
    ┌────────────────────┐
    │ Does destination   │
    │ file exist?        │
    └────┬──────────┬────┘
         │          │
      YES│          │NO
         │          │
         ▼          ▼
  ┌──────────┐  ┌──────────────┐
  │ --force  │  │ Create file  │
  │ flag?    │  │ ✅           │
  └─┬──────┬─┘  └──────────────┘
    │      │
 YES│      │NO
    │      │
    ▼      ▼
┌─────┐ ┌──────────────┐
│Over-│ │ Skip file    │
│write│ │ Return []    │
│✅   │ │ ❌           │
└─────┘ └──────────────┘

Practical Scenarios

Scenario 1: Initial Installation (Safe)

# First time deploying - no files exist yet
sudo lib_layered_config deploy \
  --source ./dist/config.toml \
  --vendor Acme --app MyApp --slug myapp \
  --target app

# ✅ Result: File created
# Output: ["/etc/xdg/myapp/config.toml"]

Scenario 2: User Has Customizations (Protected)

# User has already customized their config
# Try to deploy again without --force
lib_layered_config deploy \
  --source ./new-defaults.toml \
  --vendor Acme --app MyApp --slug myapp \
  --target user

# ❌ Result: File skipped (user's customizations preserved)
# Output: []

Scenario 3: Update During Upgrade (Intentional Overwrite)

# Major version upgrade - want to reset to new defaults
lib_layered_config deploy \
  --source ./v2-config.toml \
  --vendor Acme --app MyApp --slug myapp \
  --target user \
  --force

# ⚠️ Result: File overwritten with new version
# Output: ["/home/alice/.config/myapp/config.toml"]
# User's customizations are LOST - they should back up first!

Scenario 4: Multiple Targets (Mixed Result)

# Deploy to both app and user
# App directory is empty, user directory has existing config
lib_layered_config deploy \
  --source ./config.toml \
  --vendor Acme --app MyApp --slug myapp \
  --target app --target user

# 📋 Result: App created, user skipped
# Output: ["/etc/xdg/myapp/config.toml"]
# Note: User config not in output because it was skipped

Best Practices

✅ DO:

Test first without --force:

# See what would be deployed
lib_layered_config deploy --source ./config.toml \
  --vendor Acme --app MyApp --slug myapp --target user

# Empty output? Files exist. Check them before using --force

Use --force only when necessary:
- During clean installations
- After backing up existing configs
- When intentionally resetting to defaults

Backup before force-deploying:

# Backup user config before overwriting
cp ~/.config/myapp/config.toml ~/.config/myapp/config.toml.backup

# Now safe to force deploy
lib_layered_config deploy --source ./new-config.toml \
  --vendor Acme --app MyApp --slug myapp --target user --force

Document in installation scripts:

#!/bin/bash
# Installation script

echo "Deploying system-wide defaults..."
sudo lib_layered_config deploy \
  --source ./defaults.toml \
  --vendor Acme --app MyApp --slug myapp \
  --target app

echo "Note: User configurations preserved."
echo "To reset user config: add --force flag"

❌ DON'T:

Don't use --force in automated scripts without user confirmation:

# BAD: Might destroy user customizations
lib_layered_config deploy --source ./config.toml \
  --target user --force  # ⚠️ Dangerous!

# GOOD: Prompt user first
read -p "Overwrite existing config? (y/N): " confirm
if [ "$confirm" = "y" ]; then
  lib_layered_config deploy --source ./config.toml \
    --target user --force
fi

Don't assume empty output means failure:

# Check if command succeeded even with empty output
result=$(lib_layered_config deploy --source config.toml --target user)

# Empty array means files were skipped, not an error!
if [ "$result" = "[]" ]; then
  echo "Files already exist (not overwritten)"
fi

Python API Equivalent

The Python deploy_config() function has the same behavior:

from lib_layered_config import deploy_config

# Safe by default - won't overwrite
paths = deploy_config(
    source="./config.toml",
    vendor="Acme",
    app="MyApp",
    targets=["user"],
    slug="myapp",
    force=False  # Default
)

if not paths:
    print("File already exists and was not overwritten")
    print("Use force=True to overwrite")
else:
    print(f"Deployed to: {paths}")

# Force overwrite
paths = deploy_config(
    source="./config.toml",
    vendor="Acme",
    app="MyApp",
    targets=["user"],
    slug="myapp",
    force=True  # Overwrites existing files
)

Examples:

Example 1: Deploy system-wide defaults during installation

# Deploy app defaults to the system directory (requires sudo on Linux/macOS)
sudo lib_layered_config deploy \
  --source ./dist/config.toml \
  --vendor Acme --app MyApp --slug myapp \
  --target app

Output:

["/etc/xdg/myapp/config.toml"]

Explanation: This copies your configuration file to the system-wide location (/etc/xdg/myapp/config.toml on Linux, /Library/Application Support/Acme/MyApp/config.toml on macOS, etc.). This is typically done during package installation.

Example 2: Deploy user-specific configuration

# Deploy user config (no sudo needed)
lib_layered_config deploy \
  --source ./my-preferences.toml \
  --vendor Acme --app MyApp --slug myapp \
  --target user

Output:

["/home/alice/.config/myapp/config.toml"]

Explanation: Deploys configuration to the current user's config directory. Great for user onboarding or preference templates.

Example 3: Deploy to multiple layers

# Deploy base configuration to both system and user levels
lib_layered_config deploy \
  --source ./config/base.toml \
  --vendor Acme --app MyApp --slug myapp \
  --target app --target user \
  --force

Output:

[
  "/etc/xdg/myapp/config.toml",
  "/home/alice/.config/myapp/config.toml"
]

Explanation: Using multiple --target flags deploys the same file to multiple locations. The --force flag overwrites existing files.

Example 4: Cross-platform deployment

# Deploy for Windows even when running on Linux (for CI/testing)
lib_layered_config deploy \
  --source ./config.toml \
  --vendor Acme --app MyApp --slug myapp \
  --target user \
  --platform windows

Output:

["C:\\Users\\alice\\AppData\\Roaming\\Acme\\MyApp\\config.toml"]

Explanation: Use --platform to override platform detection. Useful for testing deployment paths on different platforms without actually being on that platform.

Example 5: Deploy host-specific configuration

# Deploy configuration specific to this server
hostname=$(hostname)
lib_layered_config deploy \
  --source ./hosts/${hostname}.toml \
  --vendor Acme --app MyApp --slug myapp \
  --target host

Output:

["/etc/xdg/myapp/hosts/server-01.toml"]

Explanation: Host-specific configurations are stored in the hosts/ subdirectory with the hostname as the filename. They override app defaults but only on machines with matching hostnames.

Example 6: Safe deployment (check before overwriting)

# Try to deploy without --force to prevent accidental overwrites
lib_layered_config deploy \
  --source ./new-config.toml \
  --vendor Acme --app MyApp --slug myapp \
  --target user

# If file exists, you'll get an empty array (nothing deployed)
# Output: []

# Then deploy with --force if you really want to overwrite
lib_layered_config deploy \
  --source ./new-config.toml \
  --vendor Acme --app MyApp --slug myapp \
  --target user \
  --force

Explanation: Without --force, the command skips existing files. This prevents accidental overwrites of user customizations.

`generate-examples`

Generate example configuration trees for documentation or onboarding.

Usage:

lib_layered_config generate-examples --destination ./examples \
  --vendor Acme --app ConfigKit --slug config-kit \
  [--platform posix|windows] \
  [--force | --no-force]

Parameters:

Parameter	Type	Required	Default	Description
`--destination`	path	Yes	-	Directory that will receive the example tree. Will be created if it doesn't exist
`--slug`	string	Yes	-	Configuration slug used in generated files
`--vendor`	string	Yes	-	Vendor namespace interpolated into examples
`--app`	string	Yes	-	Application name interpolated into examples
`--platform`	choice	No	auto-detect	Override platform layout. Valid values: `posix` (Linux/macOS layout), `windows` (Windows layout)
`--force` / `--no-force`	flag	No	`--no-force`	Overwrite existing example files

Returns: JSON array of file paths created.

Examples:

Example 1: Generate examples for your project documentation

# Create example configuration files in your docs directory
lib_layered_config generate-examples \
  --destination ./docs/configuration-examples \
  --vendor Acme --app MyApp --slug myapp

Output:

[
  "/path/to/docs/configuration-examples/xdg/myapp/config.toml",
  "/path/to/docs/configuration-examples/xdg/myapp/hosts/your-hostname.toml",
  "/path/to/docs/configuration-examples/xdg/myapp/config.d/10-override.toml",
  "/path/to/docs/configuration-examples/home/myapp/config.toml",
  "/path/to/docs/configuration-examples/.env.example"
]

File contents preview:

# docs/configuration-examples/xdg/myapp/config.toml
# Application-wide defaults for myapp
[service]
endpoint = "https://api.example.com"
timeout = 10

Explanation: Creates a complete set of example configuration files showing users how to configure your application. Include these in your documentation or repository.

Example 2: Generate both POSIX and Windows examples

# Generate Linux/macOS examples
lib_layered_config generate-examples \
  --destination ./docs/examples/unix \
  --vendor Acme --app MyApp --slug myapp \
  --platform posix

# Generate Windows examples
lib_layered_config generate-examples \
  --destination ./docs/examples/windows \
  --vendor Acme --app MyApp --slug myapp \
  --platform windows

Explanation: Generate platform-specific examples for comprehensive documentation. Windows examples use paths like ProgramData\Acme\MyApp\config.toml, while POSIX examples use /etc/xdg/myapp/config.toml.

Example 3: Update examples after configuration changes

# Regenerate examples with --force to update them
lib_layered_config generate-examples \
  --destination ./examples \
  --vendor Acme --app MyApp --slug myapp \
  --force

Explanation: When you update your configuration schema, use --force to regenerate all example files. This ensures your documentation stays in sync with your application.

Example 4: Generated file structure (POSIX)

lib_layered_config generate-examples \
  --destination ./examples \
  --vendor Acme --app MyApp --slug myapp \
  --platform posix

# View the generated structure
tree ./examples

Output:

./examples/
├── etc/
│   └── myapp/
│       ├── config.toml                    # System-wide defaults
│       └── hosts/
│           └── your-hostname.toml          # Host-specific overrides
├── xdg/
│   └── myapp/
│       ├── config.toml                    # User preferences
│       └── config.d/
│           └── 10-override.toml           # Split configuration
└── .env.example                            # Environment variable template

Explanation: The generated structure mirrors the actual configuration layout your application will use, making it easy for users to understand where to place their config files.

Example 5: Use examples as onboarding templates

# Generate examples in a temp directory
lib_layered_config generate-examples \
  --destination /tmp/myapp-examples \
  --vendor Acme --app MyApp --slug myapp

# User can copy these to actual locations
echo "To get started, copy these examples:"
echo "  sudo cp /tmp/myapp-examples/etc/myapp/config.toml /etc/myapp/"
echo "  cp /tmp/myapp-examples/xdg/myapp/config.toml ~/.config/myapp/"
echo "  cp /tmp/myapp-examples/.env.example .env"

Explanation: Generate examples in a temporary location, then provide instructions for users to copy them to the actual configuration directories.

Example 6: CI/CD - Validate configuration structure

#!/bin/bash
# In your CI pipeline, generate examples and validate them

# Generate examples
lib_layered_config generate-examples \
  --destination ./ci-examples \
  --vendor Acme --app MyApp --slug myapp

# Check that all expected files were created
expected_files=(
  "etc/myapp/config.toml"
  "xdg/myapp/config.toml"
  ".env.example"
)

for file in "${expected_files[@]}"; do
  if [ ! -f "./ci-examples/$file" ]; then
    echo "ERROR: Missing example file: $file"
    exit 1
  fi
done

echo "✓ All configuration examples are valid"

Explanation: Use in CI/CD to ensure your configuration structure is correct and all example files can be generated successfully.

`env-prefix`

Compute the canonical environment variable prefix for a configuration slug.

Usage:

lib_layered_config env-prefix <slug>

Parameters:

Parameter	Type	Required	Default	Description
`slug`	string	Yes (positional)	-	Configuration slug to convert to environment prefix

Returns: Uppercase environment prefix with dashes converted to underscores.

Examples:

Example 1: Check what environment prefix your app uses

lib_layered_config env-prefix myapp

Output:

MYAPP___

Explanation: This shows the environment variable prefix for your application (including the ___ separator). Use this prefix with double underscores for nested keys: MYAPP___DATABASE__HOST, MYAPP___SERVICE__TIMEOUT.

Example 2: Generate documentation for users

#!/bin/bash
# Script to document environment variables

app_slug="myapp"
prefix=$(lib_layered_config env-prefix "$app_slug")

cat << EOF
Environment Variables for $app_slug
====================================

All environment variables must be prefixed with: ${prefix}

Examples:
  ${prefix}DATABASE__HOST=localhost
  ${prefix}DATABASE__PORT=5432
  ${prefix}SERVICE__TIMEOUT=30
  ${prefix}SERVICE__RETRY__MAX_ATTEMPTS=3

Note: Use double underscores (__) to denote nesting in configuration keys.
EOF

Output:

Environment Variables for myapp
====================================

All environment variables must be prefixed with: MYAPP___

Examples:
  MYAPP___DATABASE__HOST=localhost
  MYAPP___DATABASE__PORT=5432
  MYAPP___SERVICE__TIMEOUT=30
  MYAPP___SERVICE__RETRY__MAX_ATTEMPTS=3

Note: Use double underscores (__) to denote nesting in configuration keys.

Explanation: Use this in documentation generation scripts to automatically show users the correct environment variable format.

Example 3: Validate environment variables in a script

#!/bin/bash
# Validate that users have set required environment variables

app_slug="config-kit"
prefix=$(lib_layered_config env-prefix "$app_slug")

required_vars=(
  "${prefix}DATABASE__HOST"
  "${prefix}DATABASE__PASSWORD"
  "${prefix}API__SECRET_KEY"
)

missing=()
for var in "${required_vars[@]}"; do
  if [ -z "${!var}" ]; then
    missing+=("$var")
  fi
done

if [ ${#missing[@]} -gt 0 ]; then
  echo "Error: Missing required environment variables:"
  printf '  %s\n' "${missing[@]}"
  exit 1
fi

echo "✓ All required environment variables are set"

Explanation: Programmatically check that required environment variables are set with the correct prefix before starting your application.

Example 4: Set test environment variables

# In a test script, set environment variables with the correct prefix
prefix=$(lib_layered_config env-prefix myapp)

export ${prefix}DATABASE__HOST="test-db.local"
export ${prefix}DATABASE__PORT="5432"
export ${prefix}SERVICE__TIMEOUT="5"

# Run tests
python -m pytest tests/

Explanation: Dynamically generate environment variable names for testing, ensuring they match your application's expected prefix.

`info`

Print package metadata including version, author, and license.

Usage:

lib_layered_config info

Parameters: None

Example:

lib_layered_config info

`fail`

Intentionally raise a RuntimeError for testing error handling and CLI behavior.

Usage:

lib_layered_config fail

Parameters: None

Raises: RuntimeError with message "i should fail".

Example:

lib_layered_config fail
# Output: RuntimeError: i should fail
# Exit code: 1

Python API

from lib_layered_config import (
    Config,
    Layer,
    read_config,
    read_config_json,
    read_config_raw,
    default_env_prefix,
    deploy_config,
    generate_examples,
    i_should_fail,
)

`Layer` Enum

The Layer enum provides type-safe constants for configuration layer names:

from lib_layered_config import Layer

# Available layers (in precedence order, lowest to highest):
Layer.DEFAULTS  # "defaults" - bundled application defaults
Layer.APP       # "app" - system-wide application config
Layer.HOST      # "host" - machine-specific overrides
Layer.USER      # "user" - per-user preferences
Layer.DOTENV    # "dotenv" - project-local .env file
Layer.ENV       # "env" - environment variables (highest precedence)

# Layer values are strings, so they work seamlessly with provenance:
origin = config.origin("service.timeout")
if origin and origin["layer"] == Layer.ENV:
    print("Value comes from environment variable")

`Config` Class

Immutable configuration value object with provenance tracking and dotted-path lookups.

Methods

`Config.get(key, default=None)`

Return the value for a dotted key path or a default when the path is missing.

Parameters:

key (str, required): Dotted path identifying nested entries (e.g., "service.timeout" or "db.host").
default (Any, optional): Value to return when the path does not resolve or encounters a non-mapping. Default: None.

Returns: The resolved value or default when missing.

Examples:

Example 1: Basic dotted-path lookup

from lib_layered_config import read_config

# Load configuration
config = read_config(vendor="Acme", app="Demo", slug="demo")

# Access nested configuration values using dotted paths
timeout = config.get("service.timeout", default=30)
endpoint = config.get("service.endpoint")
db_host = config.get("database.host", default="localhost")

print(f"Service timeout: {timeout}s")
print(f"Service endpoint: {endpoint}")
print(f"Database host: {db_host}")

Explanation: The get method traverses nested dictionaries using dot notation. If service.timeout exists in your configuration, it returns that value; otherwise, it returns the default (30).

Example 2: Handling missing keys gracefully

# This returns None if the key doesn't exist
api_key = config.get("api.secret_key")
if api_key is None:
    print("Warning: API key not configured")

# This returns a default value
max_retries = config.get("api.max_retries", default=3)
print(f"Max retries: {max_retries}")

Explanation: When you don't provide a default, get returns None for missing keys. This is useful for optional configuration values where you need to check if they were configured.

Example 3: Deep nested paths

# Access deeply nested configuration
smtp_host = config.get("email.smtp.host", default="smtp.gmail.com")
smtp_port = config.get("email.smtp.port", default=587)
use_tls = config.get("email.smtp.tls.enabled", default=True)

print(f"SMTP: {smtp_host}:{smtp_port} (TLS: {use_tls})")

Explanation: The dotted path can be arbitrarily deep. If any intermediate key is missing or not a dictionary, the default value is returned.

`Config.origin(key)`

Return provenance metadata for a dotted key when available.

Parameters:

key (str, required): Dotted key in the metadata map (e.g., "service.timeout").

Returns: Dictionary with keys layer (str), path (str | None), and key (str), or None if the key was never observed.

Examples:

Example 1: Check where a value came from

from lib_layered_config import read_config

config = read_config(vendor="Acme", app="Demo", slug="demo")

# Get provenance information
timeout_origin = config.origin("service.timeout")
if timeout_origin:
    print(f"service.timeout = {config.get('service.timeout')}")
    print(f"  Layer: {timeout_origin['layer']}")
    print(f"  Source: {timeout_origin['path'] or 'environment variable'}")
    print(f"  Key: {timeout_origin['key']}")

# Output example:
# service.timeout = 30
#   Layer: env
#   Source: environment variable
#   Key: service.timeout

Explanation: The origin method tells you which configuration layer provided a value. This is crucial for debugging when you need to understand why a particular value is being used.

Example 2: Debugging configuration precedence

# Check multiple values to understand the configuration hierarchy
keys_to_check = ["database.host", "database.port", "service.timeout"]

for key in keys_to_check:
    value = config.get(key)
    origin = config.origin(key)

    if origin:
        layer = origin['layer']
        source = origin['path'] or '(ephemeral)'
        print(f"{key}: {value} [from {layer}] {source}")
    else:
        print(f"{key}: Not configured")

# Output example:
# database.host: localhost [from user] /home/alice/.config/demo/config.toml
# database.port: 5432 [from app] /etc/demo/config.toml
# service.timeout: 30 [from env] (ephemeral)

Explanation: This shows how to audit your entire configuration to see which layer each value came from. Useful when troubleshooting unexpected configuration values.

Example 3: Validate configuration source for security

# Ensure sensitive values come from environment or dotenv
sensitive_keys = ["api.secret_key", "database.password"]

for key in sensitive_keys:
    origin = config.origin(key)
    if origin:
        if origin['layer'] not in ['env', 'dotenv']:
            print(f"WARNING: {key} should come from env/dotenv, not {origin['layer']}")
            print(f"  Currently in: {origin['path']}")
    else:
        print(f"ERROR: {key} is not configured!")

# This helps ensure secrets aren't committed to config files

Explanation: You can use provenance to enforce security policies, ensuring sensitive values only come from appropriate sources (environment variables or .env files, not checked-in config files).

`Config.as_dict()`

Return a deep, mutable copy of the configuration tree.

Parameters: None

Returns: Dictionary containing a deep copy of all configuration data.

Examples:

Example 1: Export configuration for serialization

from lib_layered_config import read_config
import json

config = read_config(vendor="Acme", app="Demo", slug="demo")

# Get a mutable copy of the entire configuration
data = config.as_dict()

# Now you can serialize it however you want
with open("config-snapshot.json", "w") as f:
    json.dump(data, f, indent=2)

print("Configuration exported to config-snapshot.json")

Explanation: Use as_dict() when you need to export or serialize the configuration data. The returned dictionary is completely independent from the original Config object.

Example 2: Modify configuration copy for testing

# Create a modified copy for testing without affecting the original
test_config = config.as_dict()
test_config["database"]["host"] = "test-db.example.com"
test_config["service"]["timeout"] = 1  # Short timeout for tests

# Original config is unchanged
print(f"Original DB: {config.get('database.host')}")  # localhost
print(f"Test DB: {test_config['database']['host']}")  # test-db.example.com

Explanation: This is useful in tests where you want to create variations of your configuration without modifying the immutable Config object.

`Config.to_json(indent=None)`

Serialize the configuration as JSON.

Parameters:

indent (int | None, optional): Indentation level for pretty-printing. None produces compact output. Default: None.

Returns: JSON string containing the configuration data.

Examples:

Example 1: Pretty-printed JSON for logs

from lib_layered_config import read_config

config = read_config(vendor="Acme", app="Demo", slug="demo")

# Pretty-printed JSON with 2-space indentation
pretty_json = config.to_json(indent=2)
print("Current configuration:")
print(pretty_json)

# Output:
# {
#   "service": {
#     "timeout": 30,
#     "endpoint": "https://api.example.com"
#   },
#   "database": {
#     "host": "localhost"
#   }
# }

Explanation: Use indent=2 or indent=4 for human-readable JSON output, perfect for logging or debugging.

Example 2: Compact JSON for APIs or storage

# Compact JSON (no whitespace)
compact_json = config.to_json()
print(compact_json)
# Output: {"service":{"timeout":30,"endpoint":"https://api.example.com"},...}

# This is useful when sending config over the network or storing in databases

Explanation: Compact JSON (no indent) minimizes the payload size, useful for network transmission or storage.

`Config.with_overrides(overrides)`

Return a new configuration with shallow top-level overrides applied.

Parameters:

overrides (Mapping[str, Any], required): Dictionary of top-level keys and values to override.

Returns: New Config instance with overrides applied, sharing provenance with the original.

Examples:

Example 1: Override configuration for specific environment

from lib_layered_config import read_config

# Load base configuration
config = read_config(vendor="Acme", app="Demo", slug="demo")

# Create a version with production overrides
prod_config = config.with_overrides({
    "service": {
        "endpoint": "https://prod-api.example.com",
        "timeout": 60
    },
    "database": {
        "host": "prod-db.example.com",
        "pool_size": 100
    }
})

print(f"Dev endpoint: {config.get('service.endpoint')}")
print(f"Prod endpoint: {prod_config.get('service.endpoint')}")

# Original config is unchanged

Explanation: This allows you to create environment-specific configurations from a base configuration without mutating the original.

Example 2: Testing with feature flags

# Enable feature flags for testing
test_config = config.with_overrides({
    "features": {
        "new_ui": True,
        "experimental_api": True,
        "debug_mode": True
    }
})

# Use test_config in your tests
if test_config.get("features.new_ui"):
    print("Running tests with new UI enabled")

Explanation: Great for testing different configurations or feature flag combinations without modifying files or environment variables.

`Config[key]` (item access)

Access top-level keys directly using bracket notation.

Parameters:

key (str): Top-level key to retrieve.

Returns: Stored value.

Raises: KeyError when key does not exist.

Examples:

Example 1: Direct access to top-level keys

from lib_layered_config import read_config

config = read_config(vendor="Acme", app="Demo", slug="demo")

# Access top-level sections directly
service_config = config["service"]
database_config = config["database"]

print(f"Service section: {service_config}")
# Output: {'timeout': 30, 'endpoint': 'https://api.example.com'}

print(f"DB host: {database_config['host']}")
# Output: localhost

Explanation: Use bracket notation config[key] to access top-level configuration sections. This returns the full nested dictionary for that section.

Example 2: Iterate over configuration sections

# Iterate over all top-level configuration keys
for section in config:
    print(f"Section: {section}")
    print(f"  Keys: {list(config[section].keys())}")

# Output:
# Section: service
#   Keys: ['timeout', 'endpoint']
# Section: database
#   Keys: ['host', 'port']

Explanation: Since Config implements the Mapping protocol, you can iterate over it like a dictionary to discover all configured sections.

`read_config`

Load and merge all configuration layers into an immutable Config object with provenance metadata.

Parameters:

vendor (str, required): Vendor namespace used to compute filesystem paths (e.g., "Acme").
app (str, required): Application name used to compute filesystem paths (e.g., "ConfigKit").
slug (str, required): Configuration slug used for file paths and environment variable prefix (e.g., "config-kit").
prefer (Sequence[str] | None, optional): Ordered sequence of preferred file suffixes (e.g., ["toml", "json", "yaml"]). Files matching earlier suffixes take precedence. Default: None (accepts all supported formats with default ordering).
start_dir (str | Path | None, optional): Starting directory for upward .env file search. Default: None (uses current working directory).
default_file (str | Path | None, optional): Path to a file injected as the lowest-precedence layer (loaded before app/host/user layers). Default: None (no defaults layer).

Returns: Immutable Config object with merged configuration and provenance tracking.

Examples:

Example 1: Basic usage - Load configuration with defaults

from lib_layered_config import read_config

# Simplest usage - just specify your app identity
config = read_config(
    vendor="Acme",
    app="MyApp",
    slug="myapp"
)

# Access configuration values
timeout = config.get("service.timeout", default=30)
endpoint = config.get("service.endpoint", default="https://api.example.com")

print(f"Service will connect to {endpoint} with {timeout}s timeout")

Explanation: This is the minimal setup. The library will automatically look for configuration files in standard locations (/etc/myapp/, ~/.config/myapp/, etc.) and merge them with environment variables.

Example 2: Using file format preferences

from lib_layered_config import read_config

# Prefer TOML files over JSON when both exist
config = read_config(
    vendor="Acme",
    app="MyApp",
    slug="myapp",
    prefer=["toml", "json", "yaml"]
)

# If both config.toml and config.json exist in the same directory,
# config.toml will be loaded because it appears first in the prefer list

Explanation: The prefer parameter controls which file format takes precedence when multiple formats exist in the same directory. This is useful when migrating from one format to another.

Example 3: Using a defaults file

from pathlib import Path
from lib_layered_config import read_config

# Start with application defaults before applying environment-specific overrides
config = read_config(
    vendor="Acme",
    app="MyApp",
    slug="myapp",
    default_file=Path("./config/defaults.toml")
)

# Precedence order now becomes:
# 1. defaults.toml (lowest)
# 2. /etc/myapp/config.toml (app layer)
# 3. /etc/myapp/hosts/hostname.toml (host layer)
# 4. ~/.config/myapp/config.toml (user layer)
# 5. .env files (dotenv layer)
# 6. Environment variables (highest)

Explanation: Use default_file to ship reasonable defaults with your application that can be overridden by system admins (app layer), per-machine configs (host layer), or users.

Example 4: Project-specific .env search

from pathlib import Path
from lib_layered_config import read_config

# Specify where to start searching for .env files
project_root = Path(__file__).parent.parent
config = read_config(
    vendor="Acme",
    app="MyApp",
    slug="myapp",
    start_dir=str(project_root)
)

# The library will search for .env files starting from project_root
# and moving upward through parent directories

Explanation: Use start_dir to control where .env file discovery begins. This ensures your project's .env file is found even if your script runs from a subdirectory.

Example 5: Complete setup with all parameters

from pathlib import Path
from lib_layered_config import read_config

# Production-ready configuration loading
config = read_config(
    vendor="Acme",
    app="MyApp",
    slug="myapp",
    prefer=["toml", "json"],  # TOML preferred
    start_dir=Path.cwd(),      # Search .env from current directory
    default_file=Path(__file__).parent / "defaults.toml"  # Ship defaults
)

# Use the configuration
db_host = config.get("database.host", default="localhost")
db_port = config.get("database.port", default=5432)
db_name = config.get("database.name", default="myapp")

print(f"Connecting to PostgreSQL at {db_host}:{db_port}/{db_name}")

# Check where each value came from
for key in ["database.host", "database.port", "database.name"]:
    origin = config.origin(key)
    if origin:
        print(f"  {key}: from {origin['layer']} layer")

Explanation: This complete example shows production-ready configuration loading with defaults, format preferences, and provenance tracking for debugging.

`read_config_json`

Load configuration and return it as JSON with provenance metadata.

Parameters:

vendor (str, required): Vendor namespace.
app (str, required): Application name.
slug (str, required): Configuration slug.
prefer (Sequence[str] | None, optional): Ordered sequence of preferred file suffixes. Default: None.
start_dir (str | Path | None, optional): Starting directory for .env search. Default: None.
default_file (str | Path | None, optional): Path to lowest-precedence defaults file. Default: None.
indent (int | None, optional): JSON indentation level. None for compact output. Default: None.

Returns: JSON string containing {"config": {...}, "provenance": {...}}.

Examples:

Example 1: API endpoint - Return configuration as JSON

from lib_layered_config import read_config_json
from flask import Flask, jsonify

app = Flask(__name__)

@app.route("/api/config")
def get_config():
    # Load and return configuration as JSON with provenance
    json_payload = read_config_json(
        vendor="Acme",
        app="MyApp",
        slug="myapp",
        indent=2  # Pretty-printed for readability
    )
    return json_payload, 200, {'Content-Type': 'application/json'}

# The response includes both config values and their sources

Explanation: Perfect for exposing configuration through APIs. The JSON includes provenance data so clients can see where each value came from.

Example 2: Configuration audit tool

from lib_layered_config import read_config_json
import json

# Load configuration with provenance
payload = read_config_json(
    vendor="Acme",
    app="MyApp",
    slug="myapp",
    indent=2
)

data = json.loads(payload)

# Audit where sensitive values come from
print("Configuration Audit Report")
print("=" * 50)

for key, info in data["provenance"].items():
    value = data["config"]
    # Navigate to the value using the key
    for part in key.split("."):
        value = value.get(part, {})

    print(f"\n{key}: {value}")
    print(f"  Source Layer: {info['layer']}")
    print(f"  File Path: {info['path'] or '(environment variable)'}")

Explanation: Use this for creating audit reports that show exactly where each configuration value originated from.

Example 3: Compact JSON for logging

from lib_layered_config import read_config_json
import logging

# Get compact JSON (no indentation) for structured logging
compact_json = read_config_json(
    vendor="Acme",
    app="MyApp",
    slug="myapp",
    indent=None  # Compact output
)

# Log the configuration snapshot
logging.info(f"Application started with config: {compact_json}")

Explanation: Compact JSON is ideal for log aggregation systems where you want to log the entire configuration as a single line.

`read_config_raw`

Return raw data and provenance mappings for advanced tooling.

Parameters:

vendor (str, required): Vendor namespace.
app (str, required): Application name.
slug (str, required): Configuration slug.
prefer (Sequence[str] | None, optional): Ordered sequence of preferred file suffixes. Default: None.
start_dir (str | None, optional): Starting directory for .env search. Default: None.
default_file (str | Path | None, optional): Path to lowest-precedence defaults file. Default: None.

Returns: Tuple of (data_dict, provenance_dict) where both are mutable dictionaries.

Examples:

Example 1: Template rendering with configuration

from lib_layered_config import read_config_raw
from jinja2 import Template

# Load configuration as raw dictionaries
data, provenance = read_config_raw(
    vendor="Acme",
    app="MyApp",
    slug="myapp"
)

# Use in template rendering
template = Template("""
Database Configuration:
  Host: {{ database.host }}
  Port: {{ database.port }}
  Database: {{ database.name }}

Service Configuration:
  Timeout: {{ service.timeout }}s
  Endpoint: {{ service.endpoint }}
""")

output = template.render(**data)
print(output)

Explanation: Raw dictionaries are perfect for template rendering where you need mutable data structures.

Example 2: Configuration validation

from lib_layered_config import read_config_raw

# Load configuration
data, provenance = read_config_raw(
    vendor="Acme",
    app="MyApp",
    slug="myapp"
)

# Validate required fields
required_keys = [
    ("database.host", str),
    ("database.port", int),
    ("service.timeout", int),
]

errors = []
for key, expected_type in required_keys:
    # Navigate the nested dictionary
    value = data
    for part in key.split("."):
        value = value.get(part) if isinstance(value, dict) else None
        if value is None:
            break

    if value is None:
        errors.append(f"Missing required key: {key}")
    elif not isinstance(value, expected_type):
        errors.append(f"{key} must be {expected_type.__name__}, got {type(value).__name__}")

if errors:
    print("Configuration validation errors:")
    for error in errors:
        print(f"  - {error}")
else:
    print("Configuration is valid!")

Explanation: Use read_config_raw for advanced validation or transformation where you need full control over the data structures.

Example 3: Merge with runtime overrides

from lib_layered_config import read_config_raw

# Load base configuration
data, provenance = read_config_raw(
    vendor="Acme",
    app="MyApp",
    slug="myapp"
)

# Apply runtime overrides (e.g., from command-line arguments)
if args.db_host:
    data["database"]["host"] = args.db_host
if args.debug:
    data["logging"]["level"] = "DEBUG"

# Now use the modified configuration
print(f"Final configuration: {data}")

Explanation: Raw dictionaries can be mutated, making them useful when you need to apply runtime overrides from command-line arguments or other sources.

`default_env_prefix`

Compute the canonical environment variable prefix for a slug.

Parameters:

slug (str, required): Configuration slug (e.g., "config-kit").

Returns: Uppercase environment prefix with dashes converted to underscores (e.g., "CONFIG_KIT").

Examples:

Example 1: Generate documentation for environment variables

from lib_layered_config import default_env_prefix

# Calculate the prefix for your application
slug = "myapp"
prefix = default_env_prefix(slug)

print(f"Environment Variables for {slug}:")
print(f"=" * 50)
print(f"\n{prefix}<SECTION>__<KEY>=<value>\n")
print("Examples:")
print(f"  {prefix}DATABASE__HOST=localhost")
print(f"  {prefix}DATABASE__PORT=5432")
print(f"  {prefix}SERVICE__TIMEOUT=30")
print(f"\nNote: Use double underscores (__) for nested keys")

Explanation: Use this to generate documentation showing users how to set environment variables for your application.

Example 2: Programmatically set environment variables

import os
from lib_layered_config import default_env_prefix

# Calculate prefix
prefix = default_env_prefix("myapp")

# Set environment variables programmatically (useful in tests)
os.environ[f"{prefix}DATABASE__HOST"] = "test-db.example.com"
os.environ[f"{prefix}DATABASE__PORT"] = "5432"
os.environ[f"{prefix}SERVICE__TIMEOUT"] = "5"

# Now when you load configuration, these will be picked up
from lib_layered_config import read_config
config = read_config(vendor="Acme", app="MyApp", slug="myapp")

print(f"DB Host: {config.get('database.host')}")  # test-db.example.com

Explanation: Programmatically generate environment variable names for testing or dynamic configuration.

Example 3: Validate environment variable names

import os
from lib_layered_config import default_env_prefix

slug = "myapp"
expected_prefix = default_env_prefix(slug)

# Check if environment variables are correctly namespaced
print(f"Checking environment variables for prefix: {expected_prefix}_")

mismatched = []
for key in os.environ:
    if "DATABASE" in key or "SERVICE" in key:
        if not key.startswith(expected_prefix + "_"):
            mismatched.append(key)

if mismatched:
    print("\nWarning: Found environment variables that won't be loaded:")
    for key in mismatched:
        correct_name = f"{expected_prefix}_{key}"
        print(f"  {key} should be {correct_name}")
else:
    print("All environment variables are correctly prefixed!")

Explanation: Validate that your environment variables are correctly prefixed so they'll be picked up by the configuration loader.

`deploy_config`

Copy a source configuration file into one or more layer directories.

Parameters:

source (str | Path, required): Path to the configuration file to copy.
vendor (str, required): Vendor namespace.
app (str, required): Application name.
targets (Sequence[str], required): Layer targets to deploy to. Valid values: "app", "host", "user".
slug (str | None, optional): Configuration slug. Default: None (uses app as slug).
platform (str | None, optional): Override auto-detected platform. Valid values: "linux", "darwin", "windows", or any value starting with "win". Default: None (auto-detects from current platform).
force (bool, optional): Overwrite existing files at destinations. Default: False.

Returns: List of Path objects for files created or overwritten.

Raises: FileNotFoundError if source file does not exist.

Examples:

Example 1: Deploy system-wide defaults

from lib_layered_config import deploy_config
from pathlib import Path

# Deploy app-wide defaults to the system directory
created_paths = deploy_config(
    source=Path("./config/defaults.toml"),
    vendor="Acme",
    app="MyApp",
    targets=["app"],  # Deploy to system-wide location
    slug="myapp"
)

# On Linux, this copies to: /etc/myapp/config.toml
# On macOS: /Library/Application Support/Acme/MyApp/config.toml
# On Windows: C:\ProgramData\Acme\MyApp\config.toml

for path in created_paths:
    print(f"Deployed to: {path}")

Explanation: Use the "app" target to deploy system-wide defaults that all users share. This is typically done during installation.

Example 2: Deploy user-specific configuration

from lib_layered_config import deploy_config

# Deploy user-specific configuration
created_paths = deploy_config(
    source="./my-config.toml",
    vendor="Acme",
    app="MyApp",
    targets=["user"],  # Deploy to user's config directory
    slug="myapp"
)

# On Linux, this copies to: ~/.config/myapp/config.toml
# On macOS: ~/Library/Application Support/Acme/MyApp/config.toml
# On Windows: %APPDATA%\Acme\MyApp\config.toml

print(f"User configuration deployed to: {created_paths[0]}")

Explanation: Use the "user" target to set up per-user configuration. Great for onboarding scripts or user preference templates.

Example 3: Deploy host-specific configuration

from lib_layered_config import deploy_config
import socket

# Deploy configuration specific to this host
hostname = socket.gethostname()
created_paths = deploy_config(
    source=f"./configs/{hostname}.toml",
    vendor="Acme",
    app="MyApp",
    targets=["host"],  # Deploy to host-specific location
    slug="myapp"
)

# On Linux, this copies to: /etc/myapp/hosts/{hostname}.toml
# The file will only be loaded on machines with this hostname

print(f"Host-specific config for '{hostname}' deployed to: {created_paths[0]}")

Explanation: Host-specific configurations override app defaults but are still system-wide. Useful for server-specific settings in multi-server deployments.

Example 4: Deploy to multiple layers at once

from lib_layered_config import deploy_config

# Deploy the same config to multiple layers
created_paths = deploy_config(
    source="./base-config.toml",
    vendor="Acme",
    app="MyApp",
    targets=["app", "user"],  # Deploy to both system and user directories
    slug="myapp",
    force=True  # Overwrite if already exists
)

print(f"Deployed to {len(created_paths)} locations:")
for path in created_paths:
    print(f"  - {path}")

Explanation: Deploy to multiple layers simultaneously. Useful for setting up consistent defaults across system and user levels. The force=True parameter allows overwriting existing files.

Example 5: Cross-platform deployment script

from lib_layered_config import deploy_config
import sys

# Deployment script that works across platforms
source_config = "./dist/config.toml"

print(f"Deploying configuration on {sys.platform}...")

try:
    created_paths = deploy_config(
        source=source_config,
        vendor="Acme",
        app="MyApp",
        targets=["app"],
        slug="myapp"
        # platform auto-detected
    )

    print(f"✓ Successfully deployed to {len(created_paths)} location(s)")
    for path in created_paths:
        print(f"  {path}")

except FileNotFoundError:
    print(f"✗ Error: Source file '{source_config}' not found")
    sys.exit(1)

Explanation: The function automatically detects the platform and deploys to the appropriate directories. Perfect for cross-platform installation scripts.

`generate_examples`

Generate example configuration trees for documentation or onboarding.

Parameters:

destination (str | Path, required): Directory that will receive the example tree.
slug (str, required): Configuration slug used in generated files.
vendor (str, required): Vendor namespace.
app (str, required): Application name.
force (bool, optional): Overwrite existing example files. Default: False.
platform (str | None, optional): Override platform layout. Valid values: "posix", "windows". Default: None (uses current platform).

Returns: List of Path objects for files created.

Examples:

Example 1: Generate documentation examples

from lib_layered_config import generate_examples
from pathlib import Path

# Generate example configuration files for documentation
docs_dir = Path("./docs/examples")
created_files = generate_examples(
    destination=docs_dir,
    slug="myapp",
    vendor="Acme",
    app="MyApp",
    platform="posix"  # Generate Linux/macOS examples
)

print(f"Generated {len(created_files)} example files:")
for file_path in created_files:
    relative = file_path.relative_to(docs_dir)
    print(f"  - {relative}")

# Output shows:
#   - etc/myapp/config.toml (app defaults)
#   - etc/myapp/hosts/your-hostname.toml (host overrides)
#   - xdg/myapp/config.toml (user preferences)
#   - xdg/myapp/config.d/10-override.toml (split overrides)
#   - .env.example (environment variables)

Explanation: Perfect for generating example configurations to include in your documentation or repository. Users can copy these examples to get started quickly.

Example 2: Generate Windows examples for cross-platform project

from lib_layered_config import generate_examples
from pathlib import Path

# Generate Windows-specific examples even on Linux/macOS
windows_examples = Path("./docs/examples-windows")
created_files = generate_examples(
    destination=windows_examples,
    slug="myapp",
    vendor="Acme",
    app="MyApp",
    platform="windows"  # Force Windows layout
)

print("Windows configuration examples:")
for file_path in created_files:
    print(f"  {file_path.relative_to(windows_examples)}")

# Output shows Windows paths:
#   - ProgramData/Acme/MyApp/config.toml
#   - ProgramData/Acme/MyApp/hosts/your-hostname.toml
#   - AppData/Roaming/Acme/MyApp/config.toml
#   - .env.example

Explanation: Generate platform-specific examples regardless of your current OS. Great for maintaining documentation for all supported platforms.

Example 3: Onboarding script - Generate and customize examples

from lib_layered_config import generate_examples
from pathlib import Path

def onboard_user(username: str):
    """Generate personalized configuration examples for a new user."""

    # Create user-specific examples directory
    user_examples = Path(f"/tmp/{username}-config-examples")
    user_examples.mkdir(exist_ok=True)

    # Generate example files
    created = generate_examples(
        destination=user_examples,
        slug="myapp",
        vendor="Acme",
        app="MyApp"
    )

    print(f"Generated {len(created)} example files for {username}:")

    # Customize the examples with user-specific values
    user_config = user_examples / "xdg/myapp/config.toml"
    if user_config.exists():
        content = user_config.read_text()
        # Add user-specific comment
        content = f"# Configuration for {username}\n" + content
        user_config.write_text(content)

    print(f"\nExamples generated in: {user_examples}")
    print("Copy these files to get started:")
    for f in created:
        print(f"  {f.relative_to(user_examples)}")

# Run onboarding
onboard_user("alice")

Explanation: Generate examples as part of an onboarding workflow. You can then customize the generated files programmatically before presenting them to users.

Example 4: Update examples (force overwrite)

from lib_layered_config import generate_examples
from pathlib import Path

# Regenerate examples, overwriting existing ones
examples_dir = Path("./examples")
created = generate_examples(
    destination=examples_dir,
    slug="myapp",
    vendor="Acme",
    app="MyApp",
    force=True  # Overwrite existing examples
)

print(f"Regenerated {len(created)} example files")

# This is useful when you update your configuration schema
# and need to refresh the documentation examples

Explanation: Use force=True when updating examples after schema changes. This ensures all example files reflect your latest configuration structure.

Example 5: Generate both POSIX and Windows examples

from lib_layered_config import generate_examples
from pathlib import Path

def generate_all_examples():
    """Generate examples for all platforms."""

    base_dir = Path("./docs/config-examples")

    # Generate POSIX examples
    posix_files = generate_examples(
        destination=base_dir / "linux-macos",
        slug="myapp",
        vendor="Acme",
        app="MyApp",
        platform="posix"
    )
    print(f"Generated {len(posix_files)} POSIX examples")

    # Generate Windows examples
    windows_files = generate_examples(
        destination=base_dir / "windows",
        slug="myapp",
        vendor="Acme",
        app="MyApp",
        platform="windows"
    )
    print(f"Generated {len(windows_files)} Windows examples")

    print(f"\nTotal: {len(posix_files) + len(windows_files)} example files")
    print(f"Location: {base_dir}")

generate_all_examples()

Explanation: Generate complete documentation showing users how to configure your app on any platform. This is essential for cross-platform applications.

`i_should_fail`

Intentionally raise a RuntimeError for testing error handling.

Parameters: None

Raises: RuntimeError with message "i should fail".

Example:

from lib_layered_config import i_should_fail

try:
    i_should_fail()
except RuntimeError as e:
    print(f"Caught expected error: {e}")

Example Generation & Deployment

Use the Python helpers or CLI equivalents:

from pathlib import Path
from lib_layered_config.examples import deploy_config, generate_examples

# copy one file into the system/user layers
paths = deploy_config("./myapp/config.toml", vendor="Acme", app="ConfigKit", targets=("app", "user"))

# scaffold an example tree for documentation
examples = generate_examples(Path("./examples"), slug="config-kit", vendor="Acme", app="ConfigKit")

Provenance & Observability

Every merged key stores metadata (layer, path, key).
Structured logging lives in lib_layered_config.observability (trace-aware log_debug, log_info, log_warn, log_error).
Use bind_trace_id("abc123") to correlate CLI/log events with your own tracing.

Type Conflict Warnings

When a later layer overwrites a scalar value with a mapping (or vice versa), a warning is emitted:

import logging
logging.basicConfig(level=logging.WARNING)

# If user.toml has: service = "disabled"
# And app.toml has:  [service]
#                    timeout = 30
# A WARNING log is emitted: "type_conflict" with details about the key, layers, and types involved

This helps identify configuration mismatches where a key changes from a simple value to a nested structure (or the reverse) across layers.

Further documentation

CHANGELOG — user-facing release notes.
CONTRIBUTING — guidelines for issues, pull requests, and coding style.
DEVELOPMENT — local tooling, recommended workflow, and release checklist.
Module Reference — architecture-aligned responsibilities per module.
LICENSE — MIT license text.

Development

pip install "lib_layered_config[dev]"
make test          # lint + type-check + pytest + coverage (fail-under=90%)
make build         # build wheel / sdist artifacts
make run -- --help # run the CLI via the repo entrypoint

The development extra now targets the latest stable releases of the toolchain (pytest 8.4.2, ruff 0.14.0, codecov-cli 11.2.3, etc.), so upgrading your local environment before running make is recommended.

Formatting gate: Ruff formatting runs in check mode during make test. Run ruff format . (or pre-commit run --all-files) before pushing and consider pre-commit install to keep local edits aligned.

Coverage gate: the maintained test suite must stay ≥90% (see pyproject.toml). Add targeted unit tests if you extend functionality.

Platform notes

Windows runners install pipx and uv automatically in CI; locally ensure pipx is on your PATH before running make test so the wheel verification step succeeds.
The journald prerequisite step runs only on Linux; macOS/Windows skips it, so there is no extra setup required on those platforms.

Continuous integration

The GitHub Actions workflow executes three jobs:

Test matrix (Linux/macOS/Windows, Python 3.13 + latest 3.x) running the same pipeline as make test.
pipx / uv verification to prove the built wheel installs cleanly with the common Python app launchers.
Notebook smoke test that executes notebooks/Quickstart.ipynb to keep the tutorial in sync using the native nbformat workflow (no compatibility shims required).
CLI jobs run through lib_cli_exit_tools.cli_session, ensuring the --traceback flag behaves the same locally and in automation.

Packaging-specific jobs (conda, Nix, Homebrew sync) were retired; the Python packaging metadata in pyproject.toml remains the single source of truth.

License

Project details

These details have not been verified by PyPI

Project links

Release history Release notifications | RSS feed

5.5.1

Mar 26, 2026

5.5.0

Feb 24, 2026

5.4.1

Feb 13, 2026

5.4.0

Jan 31, 2026

5.3.8

Jan 29, 2026

5.3.5

Jan 29, 2026

5.3.4

Jan 29, 2026

5.3.3

Jan 29, 2026

5.3.1

Jan 28, 2026

5.3.0

Jan 27, 2026

5.2.1

Jan 19, 2026

5.2.0

Dec 29, 2025

5.1.0

Dec 13, 2025

5.0.0

Dec 11, 2025

4.1.0

Dec 11, 2025

4.0.1

Dec 8, 2025

4.0.0

Dec 1, 2025

This version

3.0.1

Nov 30, 2025

3.0.0

Nov 25, 2025

2.0.0

Nov 20, 2025

1.1.1

Nov 11, 2025

1.1.0

Oct 13, 2025

1.0.0

Oct 9, 2025

Download files

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

Source Distribution

lib_layered_config-3.0.1.tar.gz (202.9 kB view details)

Uploaded Nov 30, 2025 Source

Built Distribution

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

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

lib_layered_config-3.0.1-py3-none-any.whl (95.0 kB view details)

Uploaded Nov 30, 2025 Python 3

File details

Details for the file lib_layered_config-3.0.1.tar.gz.

File metadata

Download URL: lib_layered_config-3.0.1.tar.gz
Upload date: Nov 30, 2025
Size: 202.9 kB
Tags: Source
Uploaded using Trusted Publishing? No
Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for lib_layered_config-3.0.1.tar.gz
Algorithm	Hash digest
SHA256	`dbbae42e277007e46fbef07c632faf8dd8b4b038f37f2d0328a8d67c7a4a9751`
MD5	`d4ba438873e6f54d2760e7bb90a6d979`
BLAKE2b-256	`fe019d74b9593591cc7642a9966678001404259aa579277d0157f3fcac4a7e9b`

See more details on using hashes here.

File details

Details for the file lib_layered_config-3.0.1-py3-none-any.whl.

File metadata

Download URL: lib_layered_config-3.0.1-py3-none-any.whl
Upload date: Nov 30, 2025
Size: 95.0 kB
Tags: Python 3
Uploaded using Trusted Publishing? No
Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for lib_layered_config-3.0.1-py3-none-any.whl
Algorithm	Hash digest
SHA256	`01f998e196170aef1607d34f13745b68fb818ff585151898ecccd69fad1030e1`
MD5	`b5865897a42555be4881f418997221e1`
BLAKE2b-256	`3068a5e0701b55aa42b2227c64802148b89c5e30a66a8be99483af10cfc8f09d`

See more details on using hashes here.

lib-layered-config 3.0.1

Navigation

Verified details

Maintainers

Unverified details

Project links

Meta

Classifiers

Project description

lib_layered_config

Table of Contents

Key Features

Architecture Overview

Installation

Quick Start

Understanding Key Identifiers: Vendor, App, and Slug

Vendor

App

Slug (Configuration Slug)

1. Linux/UNIX Paths

2. Environment Variable Prefix

3. Cross-Platform Consistency

Slug Naming Best Practices

Complete Example: How They Work Together

Why Three Identifiers?

Quick Reference Table

Configuration File Structure

Example Configuration File (TOML)

Understanding the Structure

1. Top-Level Keys

2. Sections (Tables)

3. Nested Sections (Subtables)

4. Arrays (Lists)

5. Data Types

Complete Python Access Example

Environment Variable Mapping

JSON and YAML Equivalents

Configuration Sources & Precedence

Linux

macOS

Windows

CLI Usage

Command Summary

read

read-json

deploy

🔒 File Overwrite Behavior

Default Behavior (without --force):

With --force Flag:

Decision Flow Diagram

Practical Scenarios

Scenario 1: Initial Installation (Safe)

Scenario 2: User Has Customizations (Protected)

Scenario 3: Update During Upgrade (Intentional Overwrite)

Scenario 4: Multiple Targets (Mixed Result)

Best Practices

✅ DO:

❌ DON'T:

Python API Equivalent

generate-examples

env-prefix

info

fail

Python API

Layer Enum

Config Class

Methods

Config.get(key, default=None)

Config.origin(key)

Config.as_dict()

Config.to_json(indent=None)

Config.with_overrides(overrides)

Config[key] (item access)

read_config

read_config_json

read_config_raw

default_env_prefix

deploy_config

generate_examples

i_should_fail

`read`

`read-json`

`deploy`

Default Behavior (without `--force`):

With `--force` Flag:

`generate-examples`

`env-prefix`

`info`

`fail`

`Layer` Enum

`Config` Class

`Config.get(key, default=None)`

`Config.origin(key)`

`Config.as_dict()`

`Config.to_json(indent=None)`

`Config.with_overrides(overrides)`

`Config[key]` (item access)

`read_config`

`read_config_json`

`read_config_raw`

`default_env_prefix`

`deploy_config`

`generate_examples`

`i_should_fail`