rtac.ac_functionalities.logs

This module contains functions for logging the data between instances/tournaments as well as stats about toournaments and results.

Classes

`AbstractLogs`(scenario)	Class with all functions and loggers concerning logging and loading RTAC and tournament data.
`NewRotatingFileHandler`(filename[, mode, ...])	Overwriting logging.handlers.RotatingFileHandler in order to log to the same line in the file.
`RTACLogs`(scenario)	Class with all functions and loggers concerning logging and loading RTAC and tournament data for ReACTR implementation.

class rtac.ac_functionalities.logs.AbstractLogs(scenario: Namespace)[source]

Bases: ABC

Class with all functions and loggers concerning logging and loading RTAC and tournament data.

Parameters:: scenario (argparse.Namespace) – Namespace containing all settings for the RTAC.

general_log(message: str) → None[source]

Log message.

Parameters:: message (str) – Any message provided as a string.
Return type:: None

init_rtac_logs() → None[source]

Initializes loggers for realtime algorithm configuration data that are shared by all methods.

Return type:: None

abstract load_data() → Any[source]: Loads the data of either last logged, or first tournament.

rtac_log(rtac_data: RTACData | RTACDatapp, tourn_stats: TournamentStats) → None[source]

Logs for realtime algorithm configuration data concerning all methods.

Parameters:

rtac_data (RTACData | RTACDatapp) – Object containing data and objects necessary throughout the rtac modules.
tourn_stats (TournamentStats) – Object containing statistics about the previous tournament.

Return type:

None

scenario_log(scenario: Namespace) → None[source]

Save RTAC scenario.

Parameters:: scenario (argparse.Namespace) – Namespace containing all settings for the RTAC.
Return type:: None

class rtac.ac_functionalities.logs.NewRotatingFileHandler(filename, mode='w', maxBytes=0, backupCount=0)[source]

Bases: RotatingFileHandler

Overwriting logging.handlers.RotatingFileHandler in order to log to the same line in the file.

class rtac.ac_functionalities.logs.RTACLogs(scenario: Namespace)[source]

Bases: AbstractLogs

Class with all functions and loggers concerning logging and loading RTAC and tournament data for ReACTR implementation.

init_ranking_logs() → None[source]

Initializes loggers for data concerning ReACTR.

Return type:: None

load_data(tourn_nr: int | None = None) → tuple[dict[str, rtac.ac_functionalities.rtac_data.Configuration], dict[str, Any], dict[str, rtac.ac_functionalities.rtac_data.Configuration], int, Optional[Any]][source]

Loads data necessary for resuming the algorithm configuration from last logged state of ReACTR.

Parameters:

tourn_nr (int | None) – Either int to load logs of tournament nr. tourn_nr or None if loading tournament nr. 0 for experimental mode.

Returns:

pool : dict[str, Configuration], Configuration pool.
assessment : dict[str, Any], Scores/ Skills, confidences of logged tournament.
contender_dict : dict[str, Configuration], List of contending Configurations from logged tournament.
tourn_nr : int, Number of logged tournament.
bandit_models : dict[str, Any], All objects needed for CPPL model employment.

Return type:

tuple

parse_array(val: str) → int | float[source]

Helper function for loading logs of nd.arrays.

Parameters:: val (str) – Loaded string to decode and transform.
Returns:: Decoded and transformed form of val.
Return type:: int or float
Raises:: ValueError – If val could not be parsed.

ranking_log(pool: dict[str, rtac.ac_functionalities.rtac_data.Configuration], assessment: dict[str, Any], tourn_nr: int, contender_dict: dict[str, rtac.ac_functionalities.rtac_data.Configuration], **kwargs) → None[source]

Logs data concerning RAC method.

Parameters:

pool (dict[str, Configuration]) – Dictionary with configuration id as key and configuration as value with scenario.contenders == #items.
assessment (dict[str, Any]) – Dictionary with configuration id as key and assessment depending on the AC method used, e.g., trueskill scores, or bandit model.
tourn (nr : int) – Number of tournament after which logs are done.
contender_dict (dict[str, Configuration]) – Dictionary with configuration id as key and configuration as value: contenders of the previous tournament.
**kwargs –
Additional keyword arguments. Possible keys include:
- standard_scaler sklearn.preprocessing.StandardScaler
- min_max_scaler sklearn.preprocessing.MinMaxScaler
- pca_obj_params sklearn.decomposition.PCA

Return type:

None