Env_logger

class conformer_rl.logging.env_logger.EnvLogger(tag: str, dir: str = 'data')

Used by the agent for logging metrics produced by the environment, for example, observations, rewards, renders, etc. Supports saving data to pickle and saving molecules as .mol files.

Parameters

• tag (str) – Unique tag for identifying the logging session.

• dir (str) – Path to root directory for where logging results should be saved.

step_data

Used for storing data for every step of a single episode. Maps from strings to lists, where each index of the list corresponds to the data for that corresponding step within the episode.

Type

dict

episode_data

Used for storing information for a single episode. Used to store the self.step_data for that episode and metadata global to the entire episode.

Type

dict

cache

Used for storing data across several episodes. Maps from strings to lists, where each index of the list corresponds to an episode.

Type

dict

clear_data() → None

Resets the logger.

clear_episode() → None

Clears episode and step data.

log_step_item(key: str, val: Any) → None

Logs a single key value pair for current step.

If an existing key is found, the value is appended to the list associated with that key.

Parameters

• key (str) – the key for the data to be added.

• val (Any) – the value of the data to be added.

log_step(step_data: dict) → None

Logs each key-value pair for current step.

If an existing key is found, the value is appended to the list associated with that key.

Parameters

step_data (dict) – Contains key-value pairs to be logged.

log_episode_item(key: str, value: Any) → None

Logs a single key-value pair to the per-episode data.

log_episode(episode_data: dict) → None

Logs each key-value pair to the per-episode data.

Also adds step_data to the per-episode data with corresponding key ‘step_data’. Existing keys will be overwritten.

Parameters

episode_data (dict) – Contains key-value pairs to be logged.

save_episode(subdir: str, save_pickle: bool = True, save_molecules: bool = False, save_cache: bool = False) → None

Saves current episode_data with options for dumping to pickle file, saving data to a cache dict, and saving the rdkit molecules as .mol files. Clears the current episode and step data.

Parameters

• subdir (str) – The directory for episode data to be saved (relative to self.dir)

• save_pickle (bool) – If True, dumps episode_data as a .pickle file.

• save_molecules (bool) – If True, and ‘molecule’ key exists in step_data, dumps each molecule generated throughout the episode as a .mol file uniquely named by the step number.

• save_cache (bool) – If True, episode data is cached to self.data.

_add_to_cache(data: dict) → None

Logs each key-value pair in data to self.cache. If an existing key is found, the value is appended to the list associated with that key.

Parameters

data (dict) – contains key-value pairs to be logged.