Config Usage

Using Configuration Files

Overview

This guide demonstrates how to load and parse configuration files using Hydra's configuration management system. Neuraldecoding provides utilities to load, validate, and extract specific sections from hierarchical YAML configurations.

Configuration Parsing Function

neuraldecoding.utils.parse_verify_config

def parse_verify_config(config: DictConfig, section_name: str) -> dict

Take in a config dict, validates the config, and return the content of a specific section.

Parameters

• config : DictConfig

  Config dictionary loaded from Hydra's compose function.

• section_name : str

  The section name to extract from the config file. Supported values are 'decoder' and 'trainer'.

Returns

• dict

  The config for the specified section in the config file.

Raises

• ValueError

  If the specified section is not found in the config file, if the section has invalid structure, or if an unsupported section name is provided.

Loading Configuration Files

Example Basic Usage

To load a configuration file, use Hydra's initialize and compose functions:

from hydra import initialize, compose

with initialize(version_base=None, config_path="../../configs/parsing_test_1"):
    cfg = compose("config")

decoder_config = parse_verify_config(cfg, 'decoder')

trainer_config = parse_verify_config(cfg, 'trainer')