Module `topicnet.cooking_machine.recipes.recipe_wrapper`

Expand source code

from typing import (
    Dict,
    Tuple,
    Union,
)

from .. import Dataset
from .. import Experiment
from ..config_parser import (
    build_experiment_environment_from_yaml_config,
    KEY_DICTIONARY_FILTER_PARAMETERS,
)


recipe_template_example = """
This string should be formatted as a confing in YAML format.
If you struggle making yours, look in other recipes for guidance.
Also cooking_machine/config_parser.py docstring
provides some insight on the matter.
{field_to_fill}
"""


class BaseRecipe:
    """
    Base class to work with recipes
    """
    def __init__(self, recipe_template):
        self.recipe_template = recipe_template
        self._recipe = None

    def __str__(self):
        if self._recipe:
            return self._recipe
        else:
            return self.recipe_template

    def format_recipe(self, *args, **kwargs) -> str:
        """
        Updates `self._recipe`
        with variables specific for the dataset.
        """
        raise NotImplementedError(
            'Method needs to be specified for the recipe template'
        )

    def build_experiment_environment(
            self,
            save_path: str,
            experiment_id: str = 'default_experiment_name',
            force_separate_thread: bool = False,
    ) -> Tuple[Experiment, Dataset]:
        """
        Returns experiment and dataset instances
        needed to perform the hyperparameter tuning on the data
        according to recipe

        Parameters
        ----------
        save_path
            path to the folder to save experiment logs and models
        experiment_id
            name of the experiment folder
        force_separate_thread
            train each model in dedicated process;
            this feature helps to handle resources in Jupyter notebooks
        """
        if self._recipe is None:
            raise ValueError(
                'Recipe missing data specific parameters. '
                'Provide them with "format_recipe" method!')

        return build_experiment_environment_from_yaml_config(
            self._recipe,
            save_path=save_path,
            experiment_id=experiment_id,
            force_separate_thread=force_separate_thread,
        )

    @staticmethod
    def _format_dictionary_filter_parameters(
            parameters: Dict[Union[int, float, str, bool], Union[int, float, str, bool]],
            indent: str) -> str:

        blank_dictionary = '{}'

        if len(parameters) == 0:
            parameters_block = blank_dictionary
        else:
            parameters_block = '\n'.join([
                f'{indent}{k}: {v}'
                for k, v in parameters.items()
            ])

        return (
            KEY_DICTIONARY_FILTER_PARAMETERS
            + ':'
            + ('\n' if parameters_block != blank_dictionary else ' ')
            + parameters_block
        )

Classes

class BaseRecipe (recipe_template)

Base class to work with recipes

Expand source code

class BaseRecipe:
    """
    Base class to work with recipes
    """
    def __init__(self, recipe_template):
        self.recipe_template = recipe_template
        self._recipe = None

    def __str__(self):
        if self._recipe:
            return self._recipe
        else:
            return self.recipe_template

    def format_recipe(self, *args, **kwargs) -> str:
        """
        Updates `self._recipe`
        with variables specific for the dataset.
        """
        raise NotImplementedError(
            'Method needs to be specified for the recipe template'
        )

    def build_experiment_environment(
            self,
            save_path: str,
            experiment_id: str = 'default_experiment_name',
            force_separate_thread: bool = False,
    ) -> Tuple[Experiment, Dataset]:
        """
        Returns experiment and dataset instances
        needed to perform the hyperparameter tuning on the data
        according to recipe

        Parameters
        ----------
        save_path
            path to the folder to save experiment logs and models
        experiment_id
            name of the experiment folder
        force_separate_thread
            train each model in dedicated process;
            this feature helps to handle resources in Jupyter notebooks
        """
        if self._recipe is None:
            raise ValueError(
                'Recipe missing data specific parameters. '
                'Provide them with "format_recipe" method!')

        return build_experiment_environment_from_yaml_config(
            self._recipe,
            save_path=save_path,
            experiment_id=experiment_id,
            force_separate_thread=force_separate_thread,
        )

    @staticmethod
    def _format_dictionary_filter_parameters(
            parameters: Dict[Union[int, float, str, bool], Union[int, float, str, bool]],
            indent: str) -> str:

        blank_dictionary = '{}'

        if len(parameters) == 0:
            parameters_block = blank_dictionary
        else:
            parameters_block = '\n'.join([
                f'{indent}{k}: {v}'
                for k, v in parameters.items()
            ])

        return (
            KEY_DICTIONARY_FILTER_PARAMETERS
            + ':'
            + ('\n' if parameters_block != blank_dictionary else ' ')
            + parameters_block
        )

Subclasses

Methods

def build_experiment_environment(self, save_path: str, experiment_id: str = 'default_experiment_name', force_separate_thread: bool = False) ‑> Tuple[Experiment, Dataset]

Returns experiment and dataset instances needed to perform the hyperparameter tuning on the data according to recipe

Parameters

save_path: path to the folder to save experiment logs and models
experiment_id: name of the experiment folder
force_separate_thread: train each model in dedicated process; this feature helps to handle resources in Jupyter notebooks

Expand source code

def build_experiment_environment(
        self,
        save_path: str,
        experiment_id: str = 'default_experiment_name',
        force_separate_thread: bool = False,
) -> Tuple[Experiment, Dataset]:
    """
    Returns experiment and dataset instances
    needed to perform the hyperparameter tuning on the data
    according to recipe

    Parameters
    ----------
    save_path
        path to the folder to save experiment logs and models
    experiment_id
        name of the experiment folder
    force_separate_thread
        train each model in dedicated process;
        this feature helps to handle resources in Jupyter notebooks
    """
    if self._recipe is None:
        raise ValueError(
            'Recipe missing data specific parameters. '
            'Provide them with "format_recipe" method!')

    return build_experiment_environment_from_yaml_config(
        self._recipe,
        save_path=save_path,
        experiment_id=experiment_id,
        force_separate_thread=force_separate_thread,
    )

def format_recipe(self, *args, **kwargs) ‑> str

Updates self._recipe with variables specific for the dataset.

Expand source code

def format_recipe(self, *args, **kwargs) -> str:
    """
    Updates `self._recipe`
    with variables specific for the dataset.
    """
    raise NotImplementedError(
        'Method needs to be specified for the recipe template'
    )