Configuration System

The configuration system provides centralized management of application settings through environment variables, configuration files, and runtime configuration.

Overview

Configuration in Babylon is managed through several modules:

babylon.config.base - Base configuration class
babylon.config.defines - Game parameter definitions (GameDefines)
babylon.config.chromadb_config - ChromaDB vector database settings
babylon.config.llm_config - LLM/AI integration settings
babylon.config.logging_config - Logging configuration

Base Configuration

The BaseConfig class serves as the foundation for all configuration management.

Database Settings

Setting	Description
`DATABASE_URL`	SQLite connection string (default: `sqlite:///babylon.db`)

ChromaDB Settings

Setting	Description
`CHROMA_PERSIST_DIR`	Vector database persistence directory
`CHROMA_COLLECTION_NAME`	Default collection name
`EMBEDDING_MODEL`	Model name for embeddings

Metrics Collection

Setting	Description
`METRICS_ENABLED`	Enable/disable metrics collection
`METRICS_INTERVAL`	Collection interval in seconds
`METRICS_RETENTION_DAYS`	Data retention period

Logging Configuration

Setting	Description
`LOG_LEVEL`	Minimum logging level
`LOG_FORMAT`	Log message format
`LOG_DIR`	Log file directory

Performance Thresholds

Setting	Description
`MAX_QUERY_LATENCY_MS`	Maximum acceptable query latency
`MIN_CACHE_HIT_RATE`	Minimum acceptable cache hit rate
`MAX_MEMORY_USAGE_GB`	Maximum memory usage threshold

Usage Examples

Accessing configuration:

from babylon.config.base import BaseConfig

# Access configuration
db_url = BaseConfig.DATABASE_URL
log_level = BaseConfig.LOG_LEVEL

# Override settings
BaseConfig.METRICS_ENABLED = False

Environment Variables

The following environment variables can be used to override default settings:

# Database (SQLite)
DATABASE_URL=sqlite:///babylon.db

# ChromaDB
BABYLON_CHROMA_PERSIST_DIR=/path/to/persist
BABYLON_EMBEDDING_MODEL=all-MiniLM-L6-v2

# Metrics
BABYLON_METRICS_ENABLED=true
BABYLON_METRICS_INTERVAL=60

# Logging
BABYLON_LOG_LEVEL=INFO
BABYLON_LOG_DIR=/path/to/logs

GameDefines

All tunable game coefficients are centralized in the GameDefines Pydantic model:

from babylon.config.defines import GameDefines

defines = GameDefines()  # Load defaults from pyproject.toml
defines.economy.extraction_efficiency  # 0.8 default
defines.consciousness.drift_sensitivity_k  # Consciousness drift rate

Economy Parameters

Control imperial rent extraction and wealth flows:

Parameter	Default	Description
`extraction_efficiency`	0.8	Fraction of surplus captured as imperial rent
`tribute_rate`	0.1	Base rate of tribute extraction per tick
`wage_floor`	0.05	Minimum wage (prevents zero-wealth)
`wealth_transfer_rate`	0.15	Rate of wealth movement along edges

Consciousness Parameters

Control ideology drift and bifurcation:

Parameter	Default	Description
`drift_sensitivity_k`	0.1	How fast ideology changes per tick
`agitation_threshold`	0.3	Minimum agitation to trigger drift
`consciousness_cap`	1.0	Maximum consciousness value
`bifurcation_enabled`	True	Enable George Jackson model

Solidarity Parameters

Control SOLIDARITY edge dynamics:

Parameter	Default	Description
`decay_base`	0.95	Solidarity decay per tick (0.95 = 5% decay)
`transmission_rate`	0.1	Consciousness spread along edges
`formation_threshold`	0.3	Consciousness needed to form new edges
`min_strength`	0.05	Edges below this are pruned

Survival Parameters

Control survival calculus:

Parameter	Default	Description
`subsistence_threshold`	0.2	Wealth level for basic survival
`sigmoid_steepness`	10.0	Steepness of P(S\|A) sigmoid curve
`loss_aversion`	2.0	Kahneman-Tversky loss aversion factor
`revolution_damping`	0.5	Reduces P(S\|R) (revolution is risky)

Territory Parameters

Control carceral geography:

Parameter	Default	Description
`heat_threshold`	0.8	Heat level triggering eviction
`heat_decay`	0.1	Heat reduction per tick
`spillover_coefficient`	0.2	Heat transferred on displacement
`detention_duration`	5	Ticks in detention before incarceration
`displacement_priority`	BALANCED	Default mode (LABOR_SCARCE/BALANCED/ELIMINATION)

See Tune Simulation Parameters for usage examples and tuning workflows.

Best Practices

Use environment variables for sensitive information
Keep defaults reasonable for development environments
Document all configuration changes in commit messages
Validate configuration at startup using Pydantic validation
Use type hints for all settings