trinity.buffer.storage.sql module

目录

trinity.buffer.storage.sql module#

SQL database storage

class trinity.buffer.storage.sql.SQLStorage(config: StorageConfig)[源代码]#

基类：object

An Storage based on SQL Database.

If wrap_in_ray in StorageConfig is True, this class will be run as a Ray Actor, and provide a remote interface to the local database.

For databases that do not support multi-processing read/write (e.g. sqlite, duckdb), please set wrap_in_ray to True.

__init__(config: StorageConfig) → None[源代码]#

classmethod get_wrapper(config: StorageConfig)[源代码]#

abstractmethod write(data: List) → None[源代码]#: Write a batch of data.

abstractmethod read(batch_size: int | None = None) → List[源代码]#: Read a batch of data.

acquire() → int[源代码]#

release() → int[源代码]#

class trinity.buffer.storage.sql.SQLExperienceStorage(config: StorageConfig)[源代码]#

基类：SQLStorage

Used as trainer input.

__init__(config: StorageConfig) → None[源代码]#

write(data: List[Experience]) → None[源代码]#: Write a batch of data.

read(batch_size: int | None = None, **kwargs) → List[Experience][源代码]#: Read a batch of data.

classmethod load_from_dataset(dataset: Dataset, config: StorageConfig) → SQLExperienceStorage[源代码]#

class trinity.buffer.storage.sql.SQLTaskStorage(config: StorageConfig)[源代码]#

基类：SQLStorage

Used as explorer input.

__init__(config: StorageConfig) → None[源代码]#

write(data: List[Dict]) → None[源代码]#: Write a batch of data.

read(batch_size: int | None = None) → List[Task][源代码]#: Read a batch of data.

classmethod load_from_dataset(dataset: Dataset, config: StorageConfig) → SQLTaskStorage[源代码]#