Configuration Bootstrapping

This page documents the REvoDesign.bootstrap module, which initialises the REvoDesign configuration system using Hydra and OmegaConf. It handles copying the template configuration tree to the user's data directory, verifying and repairing the config structure, saving and reloading config files, and managing experiment and cache directories.

---

Module-Level Constants

These constants are set at import time by REvoDesign.bootstrap.__init__:

• REVODESIGN_CONFIG_FILE (str) -- Absolute path to the user's main.yaml configuration file, determined by set_REvoDesign_config_file().
• REVODESIGN_CONFIG_DIR (str) -- Directory containing all user configuration YAML files (os.path.dirname(REVODESIGN_CONFIG_FILE)).
• EXPERIMENTS_CONFIG_DIR (str) -- Directory for experiment-specific configuration files, created by experiment_config("experiments").
• CACHE_CONFIG_DIR (str) -- Directory for cached intermediate YAML files, created by experiment_config("cache"). Note: set as a module-level variable but not included in __all__; import via from REvoDesign.bootstrap import CACHE_CONFIG_DIR.

---

Functions

REvoDesign.bootstrap.set_config.set_REvoDesign_config_file(delete_user_config_tree=False)

Sets the REvoDesign configuration directory. If the main configuration file does not exist, it will be copied from the template directory. If the configuration directory exists, it will also be checked for potential issues.

Parameters:

Name	Type	Description	Default
delete_user_config_tree	bool	Whether to delete the user's configuration tree. Defaults to False.	False

Returns:

Name	Type	Description
str		The path to the REvoDesign configuration directory.

REvoDesign.bootstrap.set_config.reload_config_file(config_name='main', overrides=None, return_hydra_config=False)

Reload a configuration yaml file in a Hydra manner. As we initialize hydra w/ initialize_config_dir, which is the REVODESIGN_CONFIG_DIR at user's data directory, the config_name is supposed as a relative path of the yaml file. e.g. config_name="experiments/my_experiment" refers to experiments/my_experiment.yaml the DictConfig object can be accessed like reload_config_file(config_name="experiments/my_experiment")["experiments"]

Parameters:

Name	Type	Description	Default
config_name	str	The name of the configuration file. Defaults to "main".	'main'
overrides	list[str]	A list of overrides to apply to the configuration. Defaults to None.	None
return_hydra_config	bool	Whether to return the Hydra configuration. Defaults to False.	False

REvoDesign.bootstrap.set_config.save_configuration(new_cfg, config_name='main')

Save configuration to specified directory

Parameters:

Name	Type	Description	Default
new_cfg	DictConfig	The configuration object to be saved	required
config_name	str	Configuration file name, defaults to "main"	'main'

Returns:

Type	Description
	None

REvoDesign.bootstrap.set_config.experiment_config(name='experiments')

Create and return the path to the experiment configuration directory.

This function creates an experiment directory based on the global configuration directory and the specified name. If the directory doesn't exist, it automatically creates it, ensuring the directory exists before returning its full path.

Parameters:

Name	Type	Description	Default
name	str	The name of the experiment directory, defaults to "experiments"	'experiments'

Returns:

Name	Type	Description
str	str	The full path to the experiment directory

REvoDesign.bootstrap.set_config.set_cache_dir()

REvoDesign.bootstrap.set_config.verify_config_tree_structure(user_config_dir, template_config_dir)

Ensures every YAML file in the template tree exists in the user's config directory. Missing files are copied over, preserving the relative directory structure.

Parameters:

Name	Type	Description	Default
user_config_dir	str	Path to the user's configuration directory where files should be present	required
template_config_dir	str	Path to the template configuration directory containing reference files	required

Returns:

Type	Description
list[str]	list[str]: List of relative paths of files that were copied from template to user config directory

REvoDesign.bootstrap.set_config.enforce_config_key_structure(user_config_dir, template_config_dir)

Compares YAML files between template and user directories and replaces files whose key structure differs from the template.

Parameters:

Name	Type	Description	Default
user_config_dir	str	Path to the user configuration directory	required
template_config_dir	str	Path to the template configuration directory	required

Returns:

Type	Description
list[str]	list[str]: Returns a list of relative paths of files that were replaced

---

Classes

REvoDesign.bootstrap.set_config.ConfigConverter

A utility class to convert omegaconf.DictConfig objects to standard Python dictionaries. This conversion is done recursively to handle nested DictConfig objects.

convert(config) staticmethod

Converts an omegaconf.DictConfig object to a standard Python dictionary.

Usage

converted_dict = ConfigConverter.convert(dict_config)

:param config: The DictConfig object to convert. :return: A standard Python dictionary representation of the input DictConfig.

---

Config YAML Schema

The configuration tree lives under src/REvoDesign/config/ as the template and is copied to the user's platform data directory (platformdirs.user_data_dir("REvoDesign")/config). The directory structure is:

Top-Level Files

File	Purpose
main.yaml	Primary configuration for UI widgets, design parameters, file paths, and algorithm settings. This is the file the user edits most often.
environ.yaml	Environment variable overrides and build-time flags (e.g., paths to third-party executables). Ignored by enforce_config_key_structure to allow local customisation.
logger.yaml	Logging system configuration (log levels, handlers, formatters).
openmm.yaml	OpenMM simulation setup parameters (force fields, integrator, thermostats).
runtime.yaml	Runtime-specific configuration for plugin operation mode.

Subdirectory Presets

rfdiffusion/ -- Parameter presets for RFdiffusion, organised by use case:

• base.yaml -- Default diffusion parameters.
• enzyme.yaml -- Enzyme-specific diffusion settings.
• motif_scaffoldding.yaml -- Motif scaffolding parameters.
• partial_diffusion.yaml -- Partial diffusion configuration.
• symmetry.yaml -- Symmetric diffusion settings.

rosetta-node/ -- Rosetta execution presets for different deployment modes:

• native.yaml -- Native (local) Rosetta execution.
• docker.yaml / docker_mpi.yaml -- Docker-based execution (serial and MPI).
• wsl.yaml / wsl_mpi.yaml -- WSL-based execution (serial and MPI).
• mpi.yaml -- General MPI execution.

sidechain-solver/ -- Sidechain solver configurations:

• pippack.yaml -- PIPPack configuration for sidechain packing.
• diffpack.yaml -- DiffPack configuration for diffusion-based packing.
• dlpacker_pytorch.yaml -- DLPacker PyTorch configuration.

third_party/scorers/ -- Third-party scorer API configurations:

• openkinetics_api.yaml -- OpenKinetics REST API endpoint and authentication settings.