API Reference

This section contains the API reference for the shared utilities in src.common.

Common Utilities Package

Base Trainer

BaseTrainer(config)

Abstract base class for all baseline method trainers

Configuration Management

read_config(config_class, config_path[, ...])	Read and validate config using dataclass
make_config(config_class, config_data[, ...])	Create dataclass instance from dictionary
save_config(config, path)	Save config to TOML file
DataConfig(name[, path, preprocessing])	Base data configuration
OutputConfig([folder, save_model, ...])	Output configuration

Data Loading

WSDataLoader(config)

Abstract base class for weak supervision data loaders

Evaluation

Evaluator()

Extended evaluator with weak supervision specific metrics

Hyperparameter Tuning

tune_hyperparameters(config, output, *[, ...])

Main hyperparameter tuning function

Utilities

set_seed([seed])	Set random seed for reproducibility
get_device([device_str])	Get PyTorch device
ensure_dir(path)	Ensure directory exists
save_json(data, path)	Save data to JSON file
load_json(path)	Load data from JSON file
Timer()