babylon.config.chromadb_config

ChromaDB configuration for the Archive layer.

The Archive stores the semantic history and ideological context. It is the collective unconscious of the simulation.

Classes

ChromaDBConfig()

Configuration for ChromaDB vector database.

class babylon.config.chromadb_config.ChromaDBConfig

Bases: object

Configuration for ChromaDB vector database.

The Archive persists embeddings locally using DuckDB+Parquet. No external servers required - fully embedded, fully materialist.

BASE_DIR: Final[Path] = PosixPath('chromadb')

THEORY_COLLECTION: Final[str] = 'marxist_theory'

HISTORY_COLLECTION: Final[str] = 'game_history'

ENTITIES_COLLECTION: Final[str] = 'entity_embeddings'

BATCH_SIZE: Final[int] = 100

SEARCH_LIMIT: Final[int] = 10

DISTANCE_THRESHOLD: Final[float] = 0.4

classmethod get_settings()

Get ChromaDB Settings object for client initialization.

Uses the new ChromaDB 1.x API. Note: persist_directory is no longer passed here - it’s passed directly to PersistentClient(path=…).

Return type:

Settings

classmethod get_persist_directory()

Get the persistence directory as a string.

Return type:

str