myco.config

Stores configuration defaults for e.g. model architectures based on yaml file inputs

ArchitectureConfig

Bases: dict

Stores architecture parameters for deep learning models

Source code in myco/config/__init__.py

class ArchitectureConfig(dict):
    """Stores architecture parameters for deep learning models"""

    def __init__(self, config: Union[dict, str]):
        """Creates the architecture configuration object.

        Assumes that there's a one-level nesting of the configuration options,
            and all "architecture" attributes will be stored as attributes in the
            output ArchitectureConfig object.

        Args:
            config: the configuration options to pass to this object.
                If it's a string, it assumes a path to a yaml file.
                If it's a dictionary, it assumes the config was already read.

        Returns:
            ArchitectureConfig object with the yaml config options available as
                object attributes (so, config.block_structure)
        """
        if type(config) is str:
            configuration = read_yaml(config)
        else:
            configuration = config

        parameters = configuration["architecture"]
        super(ArchitectureConfig, self).__init__(**parameters)
        self.__dict__ = self

__init__(config)

Creates the architecture configuration object.

Assumes that there's a one-level nesting of the configuration options, and all "architecture" attributes will be stored as attributes in the output ArchitectureConfig object.

Parameters:

Name	Type	Description	Default
config	Union[dict, str]	the configuration options to pass to this object. If it's a string, it assumes a path to a yaml file. If it's a dictionary, it assumes the config was already read.	required

Returns:

Type	Description
	ArchitectureConfig object with the yaml config options available as object attributes (so, config.block_structure)

Source code in myco/config/__init__.py

def __init__(self, config: Union[dict, str]):
    """Creates the architecture configuration object.

    Assumes that there's a one-level nesting of the configuration options,
        and all "architecture" attributes will be stored as attributes in the
        output ArchitectureConfig object.

    Args:
        config: the configuration options to pass to this object.
            If it's a string, it assumes a path to a yaml file.
            If it's a dictionary, it assumes the config was already read.

    Returns:
        ArchitectureConfig object with the yaml config options available as
            object attributes (so, config.block_structure)
    """
    if type(config) is str:
        configuration = read_yaml(config)
    else:
        configuration = config

    parameters = configuration["architecture"]
    super(ArchitectureConfig, self).__init__(**parameters)
    self.__dict__ = self

SamplingConfig

Bases: dict

Stores data processing parameters for sample drawing

Source code in myco/config/__init__.py

class SamplingConfig(dict):
    """Stores data processing parameters for sample drawing"""

    def __init__(self, config: Union[dict, str]):
        """Creates the dataset configuration object.

        Args:
            config: the configuration options to pass to this object.
                If it's a string, it assumes a path to a yaml file.
                If it's a dictionary, it assumes the config was already read.

        Returns:
            DatasetConfig object with the yaml config options available as
                object attributes (so, config.nodata)
        """
        if type(config) is str:
            configuration = read_yaml(config)
        else:
            configuration = config

        super(SamplingConfig, self).__init__(**configuration)
        self.__dict__ = self

__init__(config)

Creates the dataset configuration object.

Parameters:

Name	Type	Description	Default
config	Union[dict, str]	the configuration options to pass to this object. If it's a string, it assumes a path to a yaml file. If it's a dictionary, it assumes the config was already read.	required

Returns:

Type	Description
	DatasetConfig object with the yaml config options available as object attributes (so, config.nodata)

Source code in myco/config/__init__.py

def __init__(self, config: Union[dict, str]):
    """Creates the dataset configuration object.

    Args:
        config: the configuration options to pass to this object.
            If it's a string, it assumes a path to a yaml file.
            If it's a dictionary, it assumes the config was already read.

    Returns:
        DatasetConfig object with the yaml config options available as
            object attributes (so, config.nodata)
    """
    if type(config) is str:
        configuration = read_yaml(config)
    else:
        configuration = config

    super(SamplingConfig, self).__init__(**configuration)
    self.__dict__ = self

get_best_model_path(model_path)

Constructs a file path to the best model output from an input model filepath.

Parameters:

Name	Type	Description	Default
model_path	str	file path to the default model path.	required

Returns:

Name	Type	Description
best_model_path	str	uses model_path to define where the best model will be saved.

Source code in myco/config/__init__.py

def get_best_model_path(model_path: str) -> str:
    """Constructs a file path to the best model output from an input model filepath.

    Args:
        model_path: file path to the default model path.

    Returns:
        best_model_path: uses `model_path` to define where the best model will be saved.
    """
    base, ext = os.path.splitext(model_path.rstrip(os.path.sep))
    best_model_path = f"{base}.best{ext}"
    return best_model_path

get_model_config_names()

Return a list of valid model configuration names.

Source code in myco/config/__init__.py

def get_model_config_names() -> list:
    """Return a list of valid model configuration names."""
    return list(MODEL_CONFIGS.keys())

get_model_config_path(name)

Return the path to the default model configuration for a model type.

Parameters:

Name	Type	Description	Default
name	str	the model configuration type. Get available options from get_model_config_names().	required

Returns:

Type	Description
dict	Default model training parameters as a dictionary.

Source code in myco/config/__init__.py

def get_model_config_path(name: str) -> dict:
    """Return the path to the default model configuration for a model type.

    Args:
        name: the model configuration type. Get available options from get_model_config_names().

    Returns:
        Default model training parameters as a dictionary.
    """
    assert name in get_model_config_names(), f"Invalid config type: {name}"
    return MODEL_CONFIGS[name]["path"]