Skip to content

Configuration

Configuration management for Attune AI. Configure via direct instantiation, YAML/JSON files, or environment variables.

Overview

The configuration system provides flexible options for customizing Attune AI behavior:

  • Direct instantiation: Pass parameters to AttuneConfig()
  • YAML/JSON files: Load from attune.config.yml or attune.config.json
  • Environment variables: Use ATTUNE_* prefixed variables (EMPATHY_* also accepted for backward compatibility)
  • Validation: Automatic validation on load with helpful error messages

Quick Start

Direct Configuration

from attune import AttuneConfig

config = AttuneConfig(
    user_id="user_123",
    confidence_threshold=0.75,
    persistence_enabled=True,
)

YAML Configuration

Create attune.config.yml:

user_id: "user_123"
confidence_threshold: 0.75
persistence_enabled: true
persistence_backend: "sqlite"
persistence_path: "./attune_data"
metrics_enabled: true

Load it:

from attune import load_config

config = load_config(filepath="attune.config.yml")

Environment Variables

export ATTUNE_USER_ID="user_123"
export ATTUNE_CONFIDENCE_THRESHOLD=0.75

from attune import load_config

# Automatically loads from environment
config = load_config(use_env=True)

Class Reference

Configuration Options

Core Settings

user_id (str, default: "default_user")

Unique identifier for the user or system.

Example:

from attune import AttuneConfig

config = AttuneConfig(user_id="user_123")

confidence_threshold (float, default: 0.75)

Minimum confidence score (0.0-1.0) required for predictions and suggestions.

Higher values = More conservative (fewer, higher-quality predictions). Lower values = More aggressive (more predictions, potentially lower quality).

Example:

from attune import AttuneConfig

# Conservative: Only high-confidence predictions
config = AttuneConfig(confidence_threshold=0.85)

# Aggressive: More predictions, accept lower confidence
config = AttuneConfig(confidence_threshold=0.60)

Trust Settings

trust_building_rate (float, default: 0.05)

How much trust increases on successful interactions (0.0-1.0).

Example:

from attune import AttuneConfig

# Fast trust building (+10% per success)
config = AttuneConfig(trust_building_rate=0.10)

# Slow trust building (+2% per success)
config = AttuneConfig(trust_building_rate=0.02)

trust_erosion_rate (float, default: 0.10)

How much trust decreases on failed interactions (0.0-1.0).

Example:

from attune import AttuneConfig

# Forgiving: Small trust loss on failure
config = AttuneConfig(trust_erosion_rate=0.05)

# Strict: Large trust loss on failure
config = AttuneConfig(trust_erosion_rate=0.20)

Persistence Settings

persistence_enabled (bool, default: True)

Enable saving patterns, metrics, and state to disk.

Example:

from attune import AttuneConfig

# Production: Enable persistence
config = AttuneConfig(persistence_enabled=True)

# Testing: Disable persistence
config = AttuneConfig(persistence_enabled=False)

persistence_backend (str, default: "sqlite")

Storage backend for persistence.

Options:

  • "sqlite" - SQLite database (local development)
  • "postgresql" - PostgreSQL (production)
  • "json" - JSON files (backup/export)

Example:

from attune import AttuneConfig

# Local development
config = AttuneConfig(persistence_backend="sqlite")

# Production
config = AttuneConfig(
    persistence_backend="postgresql",
    persistence_path="postgresql://user:pass@localhost/attune",
)

persistence_path (str, default: "./attune_data")

Path for storing persistence data.

Example:

from attune import AttuneConfig

# Default location
config = AttuneConfig(persistence_path="./attune_data")

# Custom location
config = AttuneConfig(persistence_path="/var/lib/attune")

Metrics Settings

metrics_enabled (bool, default: True)

Enable metrics collection for monitoring and analytics.

Example:

from attune import AttuneConfig

config = AttuneConfig(metrics_enabled=True)

metrics_path (str, default: "./metrics.db")

Path for storing metrics data.

Example:

from attune import AttuneConfig

config = AttuneConfig(metrics_path="/var/lib/attune/metrics.db")

Pattern Library Settings

pattern_library_enabled (bool, default: True)

Enable pattern discovery and learning.

Example:

from attune import AttuneConfig

# Disable for simple use cases
config = AttuneConfig(pattern_library_enabled=False)

pattern_sharing (bool, default: True)

Enable pattern sharing across multiple agents.

Example:

from attune import AttuneConfig

# Disable pattern sharing
config = AttuneConfig(pattern_sharing=False)

pattern_confidence_threshold (float, default: 0.3)

Minimum confidence for applying learned patterns.

Example:

from attune import AttuneConfig

config = AttuneConfig(pattern_confidence_threshold=0.80)

Configuration Methods

load_config()

Load configuration from file or environment.

from attune import load_config

# Load from YAML file
config = load_config(filepath="attune.config.yml")

# Load from JSON file
config = load_config(filepath="attune.config.json")

# Load from environment variables
config = load_config(use_env=True)

# Load from file with environment overrides
config = load_config(filepath="attune.config.yml", use_env=True)

to_yaml() / to_json()

Save configuration to file.

from attune import AttuneConfig

config = AttuneConfig(user_id="user_123")

# Save as YAML
config.to_yaml("attune.config.yml")

# Save as JSON
config.to_json("attune.config.json")

validate()

Validate configuration values.

from attune import AttuneConfig

config = AttuneConfig(user_id="user_123")

try:
    config.validate()
    print("Configuration valid")
except ValueError as e:
    print(f"Configuration invalid: {e}")

Configuration Patterns

Development Configuration

# attune.dev.yml
user_id: "dev_user"
confidence_threshold: 0.70
persistence_enabled: true
persistence_backend: "sqlite"
persistence_path: "./attune_data"
metrics_enabled: true

Production Configuration

# attune.prod.yml
user_id: "prod_system"
confidence_threshold: 0.80
persistence_enabled: true
persistence_backend: "postgresql"
persistence_path: "postgresql://user:pass@db.example.com/attune"
metrics_enabled: true
metrics_path: "postgresql://user:pass@db.example.com/metrics"

# Stricter trust and pattern thresholds
trust_erosion_rate: 0.15
pattern_confidence_threshold: 0.85

Testing Configuration

from attune import AttuneConfig

# For unit tests
config = AttuneConfig(
    user_id="test_user",
    persistence_enabled=False,  # Don't save during tests
    metrics_enabled=False,      # Don't collect metrics during tests
)

Environment Variable Reference

All configuration options can be set via environment variables with the ATTUNE_ prefix (EMPATHY_ prefix is also accepted for backward compatibility):

# Core settings
export ATTUNE_USER_ID="user_123"
export ATTUNE_CONFIDENCE_THRESHOLD=0.75

# Trust settings
export ATTUNE_TRUST_BUILDING_RATE=0.05
export ATTUNE_TRUST_EROSION_RATE=0.10

# Persistence settings
export ATTUNE_PERSISTENCE_ENABLED=true
export ATTUNE_PERSISTENCE_BACKEND=sqlite
export ATTUNE_PERSISTENCE_PATH=./attune_data

# Metrics settings
export ATTUNE_METRICS_ENABLED=true
export ATTUNE_METRICS_PATH=./metrics.db

# Pattern library settings
export ATTUNE_PATTERN_LIBRARY_ENABLED=true
export ATTUNE_PATTERN_SHARING=true
export ATTUNE_PATTERN_CONFIDENCE_THRESHOLD=0.30

See Also