hafnia 0.2.0__py3-none-any.whl → 0.2.1__py3-none-any.whl

cli/config.py

@@ -80,7 +80,7 @@ class Config:
     def __init__(self, config_path: Optional[Path] = None) -> None:
         self.config_path = self.resolve_config_path(config_path)
         self.config_path.parent.mkdir(parents=True, exist_ok=True)
-        self.config_data = self.load_config()
+        self.config_data = Config.load_config(self.config_path)
 
     def resolve_config_path(self, path: Optional[Path] = None) -> Path:
         if path:
@@ -111,12 +111,25 @@ class Config:
         endpoint = self.config.platform_url + PLATFORM_API_MAPPING[method]
         return endpoint
 
-    def load_config(self) -> ConfigFileSchema:
+    @staticmethod
+    def load_config(config_path: Path) -> ConfigFileSchema:
         """Load configuration from file."""
-        if not self.config_path.exists():
+
+        # Environment variables has higher priority than config file
+        HAFNIA_API_KEY = os.getenv("HAFNIA_API_KEY")
+        HAFNIA_PLATFORM_URL = os.getenv("HAFNIA_PLATFORM_URL")
+        if HAFNIA_API_KEY and HAFNIA_PLATFORM_URL:
+            HAFNIA_PROFILE_NAME = os.getenv("HAFNIA_PROFILE_NAME", "default").strip()
+            cfg = ConfigFileSchema(
+                active_profile=HAFNIA_PROFILE_NAME,
+                profiles={HAFNIA_PROFILE_NAME: ConfigSchema(platform_url=HAFNIA_PLATFORM_URL, api_key=HAFNIA_API_KEY)},
+            )
+            return cfg
+
+        if not config_path.exists():
             return ConfigFileSchema()
         try:
-            with open(self.config_path.as_posix(), "r") as f:
+            with open(config_path.as_posix(), "r") as f:
                 data = json.load(f)
             return ConfigFileSchema(**data)
         except json.JSONDecodeError:

hafnia/data/factory.py

@@ -1,20 +1,23 @@
+import os
 from pathlib import Path
+from typing import Any
 
-from hafnia.dataset.hafnia_dataset import HafniaDataset
-from hafnia.platform.datasets import download_or_get_dataset_path
+from hafnia import utils
+from hafnia.dataset.hafnia_dataset import HafniaDataset, get_or_create_dataset_path_from_recipe
 
 
-def load_dataset(dataset_name: str, force_redownload: bool = False) -> HafniaDataset:
+def load_dataset(recipe: Any, force_redownload: bool = False) -> HafniaDataset:
     """Load a dataset either from a local path or from the Hafnia platform."""
 
-    path_dataset = get_dataset_path(dataset_name, force_redownload=force_redownload)
-    dataset = HafniaDataset.read_from_path(path_dataset)
+    path_dataset = get_dataset_path(recipe, force_redownload=force_redownload)
+    dataset = HafniaDataset.from_path(path_dataset)
     return dataset
 
 
-def get_dataset_path(dataset_name: str, force_redownload: bool = False) -> Path:
-    path_dataset = download_or_get_dataset_path(
-        dataset_name=dataset_name,
-        force_redownload=force_redownload,
-    )
+def get_dataset_path(recipe: Any, force_redownload: bool = False) -> Path:
+    if utils.is_hafnia_cloud_job():
+        return Path(os.getenv("MDI_DATASET_DIR", "/opt/ml/input/data/training"))
+
+    path_dataset = get_or_create_dataset_path_from_recipe(recipe, force_redownload=force_redownload)
+
     return path_dataset

hafnia/dataset/dataset_names.py

@@ -1,11 +1,12 @@
 from enum import Enum
 from typing import List
 
+FILENAME_RECIPE_JSON = "recipe.json"
 FILENAME_DATASET_INFO = "dataset_info.json"
 FILENAME_ANNOTATIONS_JSONL = "annotations.jsonl"
 FILENAME_ANNOTATIONS_PARQUET = "annotations.parquet"
 
-DATASET_FILENAMES = [
+DATASET_FILENAMES_REQUIRED = [
     FILENAME_DATASET_INFO,
     FILENAME_ANNOTATIONS_JSONL,
     FILENAME_ANNOTATIONS_PARQUET,

hafnia/dataset/dataset_recipe/dataset_recipe.py

@@ -0,0 +1,327 @@
+from __future__ import annotations
+
+import json
+import os
+from pathlib import Path
+from typing import TYPE_CHECKING, Any, Callable, Dict, List, Optional, Union
+
+from pydantic import (
+    field_serializer,
+    field_validator,
+)
+
+from hafnia import utils
+from hafnia.dataset.dataset_recipe import recipe_transforms
+from hafnia.dataset.dataset_recipe.recipe_types import RecipeCreation, RecipeTransform, Serializable
+from hafnia.dataset.hafnia_dataset import HafniaDataset
+
+if TYPE_CHECKING:
+    from hafnia.dataset.hafnia_dataset import HafniaDataset
+
+
+class DatasetRecipe(Serializable):
+    creation: RecipeCreation
+    operations: Optional[List[RecipeTransform]] = None
+
+    def build(self) -> HafniaDataset:
+        dataset = self.creation.build()
+        if self.operations:
+            for operation in self.operations:
+                dataset = operation.build(dataset)
+        return dataset
+
+    def append_operation(self, operation: RecipeTransform) -> DatasetRecipe:
+        """Append an operation to the dataset recipe."""
+        if self.operations is None:
+            self.operations = []
+        self.operations.append(operation)
+        return self
+
+    ### Creation Methods (using the 'from_X' )###
+    @staticmethod
+    def from_name(name: str, force_redownload: bool = False, download_files: bool = True) -> DatasetRecipe:
+        creation = FromName(name=name, force_redownload=force_redownload, download_files=download_files)
+        return DatasetRecipe(creation=creation)
+
+    @staticmethod
+    def from_path(path_folder: Path, check_for_images: bool = True) -> DatasetRecipe:
+        creation = FromPath(path_folder=path_folder, check_for_images=check_for_images)
+        return DatasetRecipe(creation=creation)
+
+    @staticmethod
+    def from_merge(recipe0: DatasetRecipe, recipe1: DatasetRecipe) -> DatasetRecipe:
+        return DatasetRecipe(creation=FromMerge(recipe0=recipe0, recipe1=recipe1))
+
+    @staticmethod
+    def from_merger(recipes: List[DatasetRecipe]) -> DatasetRecipe:
+        """Create a DatasetRecipe from a list of DatasetRecipes."""
+        if not recipes:
+            raise ValueError("The list of recipes cannot be empty.")
+        if len(recipes) == 1:
+            return recipes[0]
+        creation = FromMerger(recipes=recipes)
+        return DatasetRecipe(creation=creation)
+
+    @staticmethod
+    def from_json_str(json_str: str) -> "DatasetRecipe":
+        """Deserialize from a JSON string."""
+        data = json.loads(json_str)
+        dataset_recipe = DatasetRecipe.from_dict(data)
+        if not isinstance(dataset_recipe, DatasetRecipe):
+            raise TypeError(f"Expected DatasetRecipe, got {type(dataset_recipe).__name__}.")
+        return dataset_recipe
+
+    @staticmethod
+    def from_json_file(path_json: Path) -> "DatasetRecipe":
+        json_str = path_json.read_text(encoding="utf-8")
+        return DatasetRecipe.from_json_str(json_str)
+
+    @staticmethod
+    def from_implicit_form(recipe: Any) -> DatasetRecipe:
+        """
+        Recursively convert from implicit recipe to explicit form.
+        Handles mixed implicit/explicit recipes.
+
+        Conversion rules:
+        - str: Will get a dataset by name -> DatasetRecipeFromName
+        - Path: Will get a dataset from path -> DatasetRecipeFromPath
+        - tuple: Will merge datasets specified in the tuple -> RecipeMerger
+        - list: Will define a list of transformations -> RecipeTransforms
+
+        Example: DataRecipe from dataset name:
+        ```python
+        recipe_implicit = "mnist"
+        recipe_explicit = DatasetRecipe.from_implicit_form(recipe_implicit)
+        >>> recipe_explicit
+        DatasetRecipeFromName(dataset_name='mnist', force_redownload=False)
+        ```
+
+        Example: DataRecipe from tuple (merging multiple recipes):
+        ```python
+        recipe_implicit = ("dataset1", "dataset2")
+        recipe_explicit = DatasetRecipe.from_implicit_form(recipe_implicit)
+        >>> recipe_explicit
+        RecipeMerger(
+            recipes=[
+                DatasetRecipeFromName(dataset_name='dataset1', force_redownload=False),
+                DatasetRecipeFromName(dataset_name='dataset2', force_redownload=False)
+            ]
+        )
+
+        Example: DataRecipe from list (recipe and transformations):
+        ```python
+        recipe_implicit = ["mnist", SelectSamples(n_samples=20), Shuffle(seed=123)]
+        recipe_explicit = DatasetRecipe.from_implicit_form(recipe_implicit)
+        >>> recipe_explicit
+        Transforms(
+            recipe=DatasetRecipeFromName(dataset_name='mnist', force_redownload=False),
+            transforms=[SelectSamples(n_samples=20), Shuffle(seed=123)]
+        )
+        ```
+
+        """
+        if isinstance(recipe, DatasetRecipe):  # type: ignore
+            # It is possible to do an early return if recipe is a 'DataRecipe'-type even for nested and
+            # potentially mixed recipes. If you (really) think about it, this might surprise you,
+            # as this will bypass the conversion logic for nested recipes.
+            # However, this is not a problem as 'DataRecipe' classes are also pydantic models,
+            # so if a user introduces a 'DataRecipe'-class in the recipe (in potentially
+            # some nested and mixed implicit/explicit form) it will (due to pydantic validation) force
+            # the user to specify all nested recipes to be converted to explicit form.
+            return recipe
+
+        if isinstance(recipe, str):  # str-type is convert to DatasetFromName
+            return DatasetRecipe.from_name(name=recipe)
+
+        if isinstance(recipe, Path):  # Path-type is convert to DatasetFromPath
+            return DatasetRecipe.from_path(path_folder=recipe)
+
+        if isinstance(recipe, tuple):  # tuple-type is convert to DatasetMerger
+            recipes = [DatasetRecipe.from_implicit_form(item) for item in recipe]
+            return DatasetRecipe.from_merger(recipes=recipes)
+
+        if isinstance(recipe, list):  # list-type is convert to Transforms
+            if len(recipe) == 0:
+                raise ValueError("List of recipes cannot be empty")
+
+            dataset_recipe = recipe[0]  # First element is the dataset recipe
+            loader = DatasetRecipe.from_implicit_form(dataset_recipe)
+
+            transforms = recipe[1:]  # Remaining items are transformations
+            return DatasetRecipe(creation=loader.creation, operations=transforms)
+
+        raise ValueError(f"Unsupported recipe type: {type(recipe)}")
+
+    ### Dataset Recipe Transformations ###
+    def shuffle(recipe: DatasetRecipe, seed: int = 42) -> DatasetRecipe:
+        operation = recipe_transforms.Shuffle(seed=seed)
+        recipe.append_operation(operation)
+        return recipe
+
+    def select_samples(
+        recipe: DatasetRecipe, n_samples: int, shuffle: bool = True, seed: int = 42, with_replacement: bool = False
+    ) -> DatasetRecipe:
+        operation = recipe_transforms.SelectSamples(
+            n_samples=n_samples, shuffle=shuffle, seed=seed, with_replacement=with_replacement
+        )
+        recipe.append_operation(operation)
+        return recipe
+
+    def splits_by_ratios(recipe: DatasetRecipe, split_ratios: Dict[str, float], seed: int = 42) -> DatasetRecipe:
+        operation = recipe_transforms.SplitsByRatios(split_ratios=split_ratios, seed=seed)
+        recipe.append_operation(operation)
+        return recipe
+
+    def split_into_multiple_splits(
+        recipe: DatasetRecipe, split_name: str, split_ratios: Dict[str, float]
+    ) -> DatasetRecipe:
+        operation = recipe_transforms.SplitIntoMultipleSplits(split_name=split_name, split_ratios=split_ratios)
+        recipe.append_operation(operation)
+        return recipe
+
+    def define_sample_set_by_size(recipe: DatasetRecipe, n_samples: int, seed: int = 42) -> DatasetRecipe:
+        operation = recipe_transforms.DefineSampleSetBySize(n_samples=n_samples, seed=seed)
+        recipe.append_operation(operation)
+        return recipe
+
+    ### Conversions ###
+    def as_python_code(self, keep_default_fields: bool = False, as_kwargs: bool = True) -> str:
+        str_operations = [self.creation.as_python_code(keep_default_fields=keep_default_fields, as_kwargs=as_kwargs)]
+        if self.operations:
+            for op in self.operations:
+                str_operations.append(op.as_python_code(keep_default_fields=keep_default_fields, as_kwargs=as_kwargs))
+        operations_str = ".".join(str_operations)
+        return operations_str
+
+    def as_short_name(self) -> str:
+        """Return a short name for the transforms."""
+
+        creation_name = self.creation.as_short_name()
+        if self.operations is None or len(self.operations) == 0:
+            return creation_name
+        short_names = [creation_name]
+        for operation in self.operations:
+            short_names.append(operation.as_short_name())
+        transforms_str = ",".join(short_names)
+        return f"Recipe({transforms_str})"
+
+    def as_json_str(self, indent: int = 2) -> str:
+        """Serialize the dataset recipe to a JSON string."""
+        data = self.model_dump(mode="json")
+        # data = type_as_first_key(data)
+        return json.dumps(data, indent=indent, ensure_ascii=False)
+
+    def as_json_file(self, path_json: Path, indent: int = 2) -> None:
+        """Serialize the dataset recipe to a JSON file."""
+        json_str = self.as_json_str(indent=indent)
+        path_json.write_text(json_str, encoding="utf-8")
+
+    ### Validation and Serialization ###
+    @field_validator("creation", mode="plain")
+    @classmethod
+    def validate_creation(cls, creation: Union[Dict, RecipeCreation]) -> RecipeCreation:
+        if isinstance(creation, dict):
+            creation = Serializable.from_dict(creation)  # type: ignore[assignment]
+        if not isinstance(creation, RecipeCreation):
+            raise TypeError(f"Operation must be an instance of RecipeCreation, got {type(creation).__name__}.")
+        return creation
+
+    @field_serializer("creation")
+    def serialize_creation(self, creation: RecipeCreation) -> dict:
+        return creation.model_dump()
+
+    @field_validator("operations", mode="plain")
+    @classmethod
+    def validate_operation(cls, operations: List[Union[Dict, RecipeTransform]]) -> List[RecipeTransform]:
+        if operations is None:
+            return None
+        validated_operations = []
+        for operation in operations:
+            if isinstance(operation, dict):
+                operation = Serializable.from_dict(operation)  # type: ignore[assignment]
+            if not isinstance(operation, RecipeTransform):
+                raise TypeError(f"Operation must be an instance of RecipeTransform, got {type(operation).__name__}.")
+            validated_operations.append(operation)
+        return validated_operations
+
+    @field_serializer("operations")
+    def serialize_operations(self, operations: Optional[List[RecipeTransform]]) -> Optional[List[dict]]:
+        """Serialize the operations to a list of dictionaries."""
+        if operations is None:
+            return None
+        return [operation.model_dump() for operation in operations]
+
+
+def unique_name_from_recipe(recipe: DatasetRecipe) -> str:
+    if isinstance(recipe.creation, FromName) and recipe.operations is None:
+        # If the dataset recipe is simply a DatasetFromName, we bypass the hashing logic
+        # and return the name directly. The dataset is already uniquely identified by its name.
+        # Add  version if need... Optionally, you may also completely delete this exception
+        # and always return the unique name including the hash to support versioning.
+        return recipe.creation.name  # Dataset name e.g 'mnist'
+    recipe_json_str = recipe.model_dump_json()
+    hash_recipe = utils.hash_from_string(recipe_json_str)
+    short_recipe_str = recipe.as_short_name()
+    unique_name = f"{short_recipe_str}_{hash_recipe}"
+    return unique_name
+
+
+def get_dataset_path_from_recipe(recipe: DatasetRecipe, path_datasets: Optional[Union[Path, str]] = None) -> Path:
+    path_datasets = path_datasets or utils.PATH_DATASETS
+    path_datasets = Path(path_datasets)
+    unique_dataset_name = unique_name_from_recipe(recipe)
+    return path_datasets / unique_dataset_name
+
+
+class FromPath(RecipeCreation):
+    path_folder: Path
+    check_for_images: bool = True
+
+    @staticmethod
+    def get_function() -> Callable[..., "HafniaDataset"]:
+        return HafniaDataset.from_path
+
+    def as_short_name(self) -> str:
+        return f"'{self.path_folder}'".replace(os.sep, "|")
+
+
+class FromName(RecipeCreation):
+    name: str
+    force_redownload: bool = False
+    download_files: bool = True
+
+    @staticmethod
+    def get_function() -> Callable[..., "HafniaDataset"]:
+        return HafniaDataset.from_name
+
+    def as_short_name(self) -> str:
+        return self.name
+
+
+class FromMerge(RecipeCreation):
+    recipe0: DatasetRecipe
+    recipe1: DatasetRecipe
+
+    @staticmethod
+    def get_function():
+        return HafniaDataset.merge
+
+    def as_short_name(self) -> str:
+        merger = FromMerger(recipes=[self.recipe0, self.recipe1])
+        return merger.as_short_name()
+
+
+class FromMerger(RecipeCreation):
+    recipes: List[DatasetRecipe]
+
+    def build(self) -> HafniaDataset:
+        """Build the dataset from the merged recipes."""
+        datasets = [recipe.build() for recipe in self.recipes]
+        return self.get_function()(datasets=datasets)
+
+    @staticmethod
+    def get_function():
+        return HafniaDataset.from_merger
+
+    def as_short_name(self) -> str:
+        return f"Merger({','.join(recipe.as_short_name() for recipe in self.recipes)})"

hafnia/dataset/dataset_recipe/recipe_transforms.py

@@ -0,0 +1,53 @@
+from typing import TYPE_CHECKING, Callable, Dict
+
+from hafnia.dataset.dataset_recipe.recipe_types import RecipeTransform
+from hafnia.dataset.hafnia_dataset import HafniaDataset
+
+if TYPE_CHECKING:
+    pass
+
+
+class Shuffle(RecipeTransform):
+    seed: int = 42
+
+    @staticmethod
+    def get_function() -> Callable[..., "HafniaDataset"]:
+        return HafniaDataset.shuffle
+
+
+class SelectSamples(RecipeTransform):
+    n_samples: int
+    shuffle: bool = True
+    seed: int = 42
+    with_replacement: bool = False
+
+    @staticmethod
+    def get_function() -> Callable[..., "HafniaDataset"]:
+        return HafniaDataset.select_samples
+
+
+class SplitsByRatios(RecipeTransform):
+    split_ratios: Dict[str, float]
+    seed: int = 42
+
+    @staticmethod
+    def get_function() -> Callable[..., "HafniaDataset"]:
+        return HafniaDataset.splits_by_ratios
+
+
+class SplitIntoMultipleSplits(RecipeTransform):
+    split_name: str
+    split_ratios: Dict[str, float]
+
+    @staticmethod
+    def get_function() -> Callable[..., "HafniaDataset"]:
+        return HafniaDataset.split_into_multiple_splits
+
+
+class DefineSampleSetBySize(RecipeTransform):
+    n_samples: int
+    seed: int = 42
+
+    @staticmethod
+    def get_function() -> Callable[..., "HafniaDataset"]:
+        return HafniaDataset.define_sample_set_by_size

hafnia/dataset/dataset_recipe/recipe_types.py

@@ -0,0 +1,140 @@
+from __future__ import annotations
+
+from abc import ABC, abstractmethod
+from collections.abc import Callable
+from typing import TYPE_CHECKING, Any, Dict, List
+
+from pydantic import BaseModel, computed_field
+
+from hafnia import utils
+
+if TYPE_CHECKING:
+    from hafnia.dataset.hafnia_dataset import HafniaDataset
+
+
+class Serializable(BaseModel, ABC):
+    @computed_field  # type: ignore[prop-decorator]
+    @property
+    def __type__(self) -> str:
+        return self.__class__.__name__
+
+    @classmethod
+    def get_nested_subclasses(cls) -> List[type["Serializable"]]:
+        """Recursively get all subclasses of a class."""
+        all_subclasses = []
+        for subclass in cls.__subclasses__():
+            all_subclasses.append(subclass)
+            all_subclasses.extend(subclass.get_nested_subclasses())
+        return all_subclasses
+
+    @classmethod
+    def name_to_type_mapping(cls) -> Dict[str, type["Serializable"]]:
+        """Create a mapping from class names to class types."""
+        return {subclass.__name__: subclass for subclass in cls.get_nested_subclasses()}
+
+    @staticmethod
+    def from_dict(data: Dict) -> "Serializable":
+        dataset_spec_args = data.copy()
+        dataset_type_name = dataset_spec_args.pop("__type__", None)
+        name_to_type_mapping = Serializable.name_to_type_mapping()
+        SerializableClass = name_to_type_mapping[dataset_type_name]
+        return SerializableClass(**dataset_spec_args)
+
+    def get_kwargs(self, keep_default_fields: bool) -> Dict:
+        """Return a dictionary of fields that are not set to their default values."""
+        kwargs = dict(self)
+        kwargs.pop("__type__", None)
+
+        if keep_default_fields:
+            return kwargs
+
+        kwargs_no_defaults = {}
+        for key, value in kwargs.items():
+            default_value = self.model_fields[key].get_default()
+            if value != default_value:
+                kwargs_no_defaults[key] = value
+
+        return kwargs_no_defaults
+
+    @abstractmethod
+    def as_short_name(self) -> str:
+        pass
+
+    def as_python_code(self, keep_default_fields: bool = False, as_kwargs: bool = True) -> str:
+        """Generate code representation of the operation."""
+        kwargs = self.get_kwargs(keep_default_fields=keep_default_fields)
+
+        args_as_strs = []
+        for argument_name, argument_value in kwargs.items():
+            # In case an argument is a Serializable, we want to keep its default fields
+            str_value = recursive_as_code(argument_value, keep_default_fields=keep_default_fields, as_kwargs=as_kwargs)
+            if as_kwargs:
+                args_as_strs.append(f"{argument_name}={str_value}")
+            else:
+                args_as_strs.append(str_value)
+
+        args_as_str = ", ".join(args_as_strs)
+        class_name = self.__class__.__name__
+        function_name = utils.pascal_to_snake_case(class_name)
+        return f"{function_name}({args_as_str})"
+
+
+def recursive_as_code(value: Any, keep_default_fields: bool = False, as_kwargs: bool = True) -> str:
+    if isinstance(value, Serializable):
+        return value.as_python_code(keep_default_fields=keep_default_fields, as_kwargs=as_kwargs)
+
+    elif isinstance(value, list):
+        as_strs = []
+        for item in value:
+            str_item = recursive_as_code(item, keep_default_fields=keep_default_fields, as_kwargs=as_kwargs)
+            as_strs.append(str_item)
+        as_str = ", ".join(as_strs)
+        return f"[{as_str}]"
+
+    elif isinstance(value, dict):
+        as_strs = []
+        for key, item in value.items():
+            str_item = recursive_as_code(item, keep_default_fields=keep_default_fields, as_kwargs=as_kwargs)
+            as_strs.append(f"{key!r}: {str_item}")
+        as_str = ", ".join(as_strs)
+        return "{" + as_str + "}"
+
+    return f"{value!r}"
+
+
+class RecipeCreation(Serializable):
+    @staticmethod
+    @abstractmethod
+    def get_function() -> Callable[..., "HafniaDataset"]:
+        pass
+
+    def build(self) -> "HafniaDataset":
+        from hafnia.dataset.dataset_recipe.dataset_recipe import DatasetRecipe
+
+        kwargs = dict(self)
+        kwargs_recipes_as_datasets = {}
+        for key, value in kwargs.items():
+            if isinstance(value, DatasetRecipe):
+                value = value.build()
+                key = key.replace("recipe", "dataset")
+            kwargs_recipes_as_datasets[key] = value
+        return self.get_function()(**kwargs_recipes_as_datasets)
+
+    def as_python_code(self, keep_default_fields: bool = False, as_kwargs: bool = True) -> str:
+        """Generate code representation of the operation."""
+        as_python_code = Serializable.as_python_code(self, keep_default_fields=keep_default_fields, as_kwargs=as_kwargs)
+        return f"DatasetRecipe.{as_python_code}"
+
+
+class RecipeTransform(Serializable):
+    @staticmethod
+    @abstractmethod
+    def get_function() -> Callable[..., "HafniaDataset"]:
+        pass
+
+    def build(self, dataset: "HafniaDataset") -> "HafniaDataset":
+        kwargs = dict(self)
+        return self.get_function()(dataset=dataset, **kwargs)
+
+    def as_short_name(self) -> str:
+        return self.__class__.__name__