Version: 1.0.0 Last Updated: 2026-05-14 Status: Active

bast-ai-config¶

Version: 0.6.3 Module: bast_ai_config Path: src/python/config

Overview¶

Configuration management library for Bast AI services using Hydra. Provides centralized configuration loading, CAT (Client Application Template) management, and environment-based overrides.

Features¶

Configuration Loading¶

YAML-based configuration files
Environment variable overrides
Configuration composition
Runtime configuration updates

CAT Management¶

CAT (Client Application Template) configurations define client-specific settings:

from bast_ai_config import get_bast_config

config = get_bast_config()
cat = config.get_cat("mycat")

# Access CAT properties
print(cat.name)
print(cat.mongodb_uri)
print(cat.openai_model)

Configuration Structure¶

Directory Layout¶

resources/config/
├── bast.yaml           # Global defaults
├── mycat.yaml          # Client-specific config
├── testcat.yaml        # Test configuration
└── ...

Configuration Schema¶

# resources/config/{catname}.yaml
name: mycat
description: "My Application"

# Database settings
mongodb:
  uri: ${MONGODB_URI:mongodb://localhost:27017}
  database: mycat_db

# OpenAI settings
openai:
  model: gpt-4o
  temperature: 0.7
  max_tokens: 1000

# Milvus settings
milvus:
  host: ${MILVUS_HOST:localhost}
  port: ${MILVUS_PORT:19530}
  collection: mycat_vectors

# Agent settings
agent:
  type: agno
  instructions: "You are a helpful assistant."

# Feature flags
features:
  enable_web_search: true
  enable_memory: true

Dependencies¶

Internal Dependencies¶

Package	Purpose
bast-ai-base	Core utilities

External Dependencies¶

Package	Version	Purpose
hydra-core	>=1.3	Configuration management

Usage Examples¶

Basic Configuration Access¶

from bast_ai_config import get_bast_config

# Get global config
config = get_bast_config()

# Get specific CAT
cat = config.get_cat("mycat")

# Access nested configuration
mongo_uri = cat.mongodb.uri
openai_model = cat.openai.model

Environment Overrides¶

# Override via environment variables
export MONGODB_URI="mongodb://prod-server:27017"
export MILVUS_HOST="prod-milvus"

Configuration in Services¶

from bast_ai.base import BastObject
from bast_ai_config import get_bast_config

class MyService(BastObject):
    def __init__(self, catname: str):
        BastObject.__init__(self, __name__)
        self._cat = get_bast_config().get_cat(catname)

    @property
    def mongo_uri(self) -> str:
        return self._cat.mongodb.uri

Dynamic Configuration¶

from bast_ai_config import get_bast_config

config = get_bast_config()

# Check if feature is enabled
if config.get_cat("mycat").features.enable_web_search:
    # Enable web search functionality
    pass

# Get configuration with default
timeout = getattr(cat.api, 'timeout', 30)

Testing¶

Unit Tests (1 file)¶

Basic configuration loading tests
CAT retrieval tests
Environment override tests

Running Tests¶

pants test src/python/config:: -- -s --no-header

Issues & Improvements¶

Current Issues¶

No configuration validation: Invalid configs not caught early
Limited hot-reload: Config changes require restart
No secrets management: Sensitive values in plain text

Suggested Improvements¶

Schema validation: Add Pydantic models for config validation
Hot reload: Support runtime configuration updates
Secrets integration: Support HashiCorp Vault, AWS Secrets Manager
Configuration versioning: Track config changes over time
Environment profiles: Support dev/staging/prod profiles

Best Practices¶

Configuration Organization¶

# Separate concerns
database:
  # All database settings

api:
  # All API settings

models:
  # All model settings

features:
  # All feature flags

Environment Variables¶

# Use defaults for local development
mongodb:
  uri: ${MONGODB_URI:mongodb://localhost:27017}

# Required in production (no default)
openai:
  api_key: ${OPENAI_API_KEY}

Feature Flags¶

features:
  # Boolean flags
  enable_caching: true
  enable_metrics: false

  # Percentage rollout
  new_feature_percentage: 50