♻️Dataset modernization

src/plaid/bridges/huggingface_bridge.py

@@ -42,13 +42,14 @@
 from huggingface_hub import HfApi, hf_hub_download, snapshot_download
 from pydantic import ValidationError
 
-from plaid import Dataset, ProblemDefinition, Sample
+from plaid import Dataset, ProblemDefinition, Sample, problem_definition
 from plaid.containers.features import SampleFeatures
 from plaid.types import IndexType
 from plaid.utils.cgns_helper import (
     flatten_cgns_tree,
     unflatten_cgns_tree,
 )
+from plaid.containers import FeatureIdentifier
 
 # ------------------------------------------------------------------------------
 
@@ -696,18 +697,29 @@ def huggingface_description_to_problem_definition(
     """
     description = {} if description == "" else description
     problem_definition = ProblemDefinition()
+    #todo
     for func, key in [
         (problem_definition.set_task, "task"),
         (problem_definition.set_split, "split"),
-        (problem_definition.add_input_scalars_names, "in_scalars_names"),
-        (problem_definition.add_output_scalars_names, "out_scalars_names"),
-        (problem_definition.add_input_fields_names, "in_fields_names"),
-        (problem_definition.add_output_fields_names, "out_fields_names"),
-        (problem_definition.add_input_meshes_names, "in_meshes_names"),
-        (problem_definition.add_output_meshes_names, "out_meshes_names"),
     ]:
         try:
-            func(description[key])
+            func([  description[key] ])
+        except KeyError:
+            logger.error(f"Could not retrieve key:'{key}' from description")
+            pass
+
+
+    for func, ftype, key in [
+        (problem_definition.add_in_features_identifiers, "scalar","in_scalars_names"),
+        (problem_definition.add_out_feature_identifier, "scalar","out_scalars_names"),
+        (problem_definition.add_in_features_identifiers, "field","in_fields_names"),
+        (problem_definition.add_out_feature_identifier, "field","out_fields_names"),
+        (problem_definition.add_in_features_identifiers, "mesh","in_meshes_names"),
+        (problem_definition.add_out_feature_identifier, "mesh","out_meshes_names"),
+
+    ]:
+        try:
+            func([ FeatureIdentifier({"type":ftype, "name":name }) for name in  description[key] ])
         except KeyError:
             logger.error(f"Could not retrieve key:'{key}' from description")
             pass

src/plaid/containers/dataset.py

@@ -37,7 +37,7 @@
 from plaid.containers.utils import check_features_size_homogeneity
 from plaid.types import Array, Feature
 from plaid.utils.base import DeprecatedError, ShapeError, generate_random_ASCII
-from plaid.utils.deprecation import deprecated, deprecated_argument
+from plaid.utils.deprecation import deprecated
 
 logger = logging.getLogger(__name__)
 
@@ -65,7 +65,6 @@ def _process_sample(path: Union[str, Path]) -> tuple:  # pragma: no cover
 class Dataset(object):
     """A set of samples, and optionnaly some other informations about the Dataset."""
 
-    @deprecated_argument("directory_path", "path", version="0.1.8", removal="0.2.0")
     def __init__(
         self,
         path: Optional[Union[str, Path]] = None,
@@ -705,19 +704,6 @@ def extract_dataset_from_identifier(
             dataset.add_sample(sample=extracted_sample, id=id)
         return dataset
 
-    @deprecated(
-        "`Dataset.from_features_identifier(...)` is deprecated, use instead `Dataset.extract_dataset_from_identifier(...)`",
-        version="0.1.8",
-        removal="0.2",
-    )
-    def from_features_identifier(
-        self,
-        feature_identifiers: Union[FeatureIdentifier, list[FeatureIdentifier]],
-    ) -> Self:
-        """DEPRECATED: Use :meth:`Dataset.extract_dataset_from_identifier` instead."""
-        return self.extract_dataset_from_identifier(
-            feature_identifiers
-        )  # pragma: no cover
 
     def get_tabular_from_homogeneous_identifiers(
         self,
@@ -829,22 +815,6 @@ def add_features_from_tabular(
 
         return dataset
 
-    @deprecated(
-        "`Dataset.from_tabular(...)` is deprecated, use instead `Dataset.add_features_from_tabular(...)`",
-        version="0.1.8",
-        removal="0.2",
-    )
-    def from_tabular(
-        self,
-        tabular: Array,
-        feature_identifiers: Union[FeatureIdentifier, list[FeatureIdentifier]],
-        restrict_to_features: bool = True,
-    ) -> Self:
-        """DEPRECATED: Use :meth:`Dataset.add_features_from_tabular` instead."""
-        return self.add_features_from_tabular(
-            tabular, feature_identifiers, restrict_to_features
-        )  # pragma: no cover
-
     # -------------------------------------------------------------------------#
     def add_info(self, cat_key: str, info_key: str, info: str) -> None:
         """Add information to the :class:`Dataset <plaid.containers.dataset.Dataset>`, overwriting existing information if there's a conflict.
@@ -1085,16 +1055,6 @@ def merge_dataset_by_features(cls, datasets_list: list[Self]) -> Self:
             merged_dataset = merged_dataset.merge_features(dataset, in_place=False)
         return merged_dataset
 
-    @deprecated_argument("directory_path", "path", version="0.1.8", removal="0.2.0")
-    @deprecated(
-        "`Dataset.save(...)` is deprecated, use instead `Dataset.save_to_file(...)`",
-        version="0.1.10",
-        removal="0.2.0",
-    )
-    def save(self, path: Union[str, Path]) -> None:
-        """DEPRECATED: use :meth:`Dataset.save_to_file` instead."""
-        self.save_to_file(path)
-
     def save_to_file(self, path: Union[str, Path]) -> None:
         """Saves the data set to a TAR (Tape Archive) file.
 
@@ -1372,19 +1332,6 @@ def check_feature_completeness(self) -> str:
         return report
 
     @classmethod
-    @deprecated(
-        "`Dataset.from_list_of_samples(samples)` is deprecated, use instead `Dataset(samples=samples)`",
-        version="0.1.8",
-        removal="0.2.0",
-    )
-    def from_list_of_samples(
-        cls, list_of_samples: list[Sample], ids: Optional[list[int]] = None
-    ) -> Self:
-        """DEPRECATED: use `Dataset(samples=..., sample_ids=...)` instead."""
-        return cls(samples=list_of_samples, sample_ids=ids)
-
-    @classmethod
-    @deprecated_argument("fname", "path", version="0.1.8", removal="0.2.0")
     def load_from_file(
         cls, path: Union[str, Path], verbose: bool = False, processes_number: int = 0
     ) -> Self:
@@ -1404,7 +1351,6 @@ def load_from_file(
         return instance
 
     @classmethod
-    @deprecated_argument("fname", "path", version="0.1.8", removal="0.2.0")
     def load_from_dir(
         cls,
         path: Union[str, Path],
@@ -1430,7 +1376,6 @@ def load_from_dir(
         )
         return instance
 
-    @deprecated_argument("fname", "path", version="0.1.8", removal="0.2.0")
     def load(
         self, path: Union[str, Path], verbose: bool = False, processes_number: int = 0
     ) -> None:
@@ -1479,7 +1424,6 @@ def load(
             )
 
     # -------------------------------------------------------------------------#
-    @deprecated_argument("save_dir", "path", version="0.1.8", removal="0.2.0")
     def add_to_dir(
         self,
         sample: Sample,
@@ -1538,15 +1482,6 @@ def add_to_dir(
         sample_dname = samples_dir / f"sample_{i_sample:09d}"
         sample.save_to_dir(sample_dname)
 
-    @deprecated(
-        "`Dataset._save_to_dir_(path)` is deprecated, use instead `Dataset.save_to_dir(path)`",
-        version="0.1.10",
-        removal="0.2.0",
-    )
-    def _save_to_dir_(self, path: Union[str, Path], verbose: bool = False) -> None:
-        """DEPRECATED: use :meth:`Dataset.save_to_dir` instead."""
-        self.save_to_dir(path, verbose=verbose)
-
     def _load_from_dir_(
         self,
         path: Union[str, Path],
@@ -1651,13 +1586,6 @@ def update(self, *a):
         if len(self) == 0:  # pragma: no cover
             print("Warning: dataset contains no sample")
 
-    @staticmethod
-    def _load_number_of_samples_(_path: Union[str, Path]) -> int:
-        """DEPRECATED: use :meth:`plaid.get_number_of_samples <plaid.containers.utils.get_number_of_samples>` instead."""
-        raise DeprecatedError(
-            'use instead: plaid.get_number_of_samples("path-to-my-dataset")'
-        )
-
     # -------------------------------------------------------------------------#
     def set_samples(self, samples: dict[int, Sample]) -> None:
         """Set the samples of the data set, overwriting the existing ones.

src/plaid/containers/feature_identifier.py

@@ -35,3 +35,41 @@ def __lt__(self, other: "FeatureIdentifier") -> bool:
             bool: True if this feature identifier is less than the other, False otherwise.
         """
         return sorted(self.items()) < sorted(other.items())
+
+    @classmethod
+    def from_string(cls, string_identifier: str) -> "FeatureIdentifier":
+        """Create a FeatureIdentifier from a string representation.
+
+        Args:
+            string_identifier (str): The string representation of the feature identifier.
+
+        The `string_identifier` must follow the format:
+            "<feature_type>::<detail1>/<detail2>/.../<detailN>"
+
+        Supported feature types:
+            - "scalar": expects 1 detail → ["name"],
+            - "field": up to 5 details → ["name", "location", "zone_name", "base_name", "time"],
+            - "nodes":  up to 3 details → ["zone_name", "base_name", "time"],
+
+        Returns:
+            FeatureIdentifier
+
+
+        Warnings:
+            - If "time" is present in a field/nodes identifier, it is cast to float.
+            - `name` is required for scalar and field features.
+        """
+        splitted_identifier = string_identifier.split("::")
+
+        feature_type = splitted_identifier[0]
+        feature_details = [detail for detail in splitted_identifier[1].split("/")]
+
+        from plaid.constants import AUTHORIZED_FEATURE_TYPES, AUTHORIZED_FEATURE_INFOS
+
+        assert feature_type in AUTHORIZED_FEATURE_TYPES, "feature_type not known"
+
+        arg_names = AUTHORIZED_FEATURE_INFOS[feature_type]
+        assert len(arg_names) >= len(feature_details), "Too much details provided"
+        data = {"type": feature_type}
+        data.update({ arg_name : feature_detail for arg_name, feature_detail in zip(arg_names, feature_details)})
+        return FeatureIdentifier(data)

src/plaid/containers/features.py

@@ -22,7 +22,6 @@
 )
 from plaid.types import Array, CGNSNode, CGNSTree, Field
 from plaid.utils import cgns_helper as CGH
-from plaid.utils.deprecation import deprecated
 
 logger = logging.getLogger(__name__)
 
@@ -129,15 +128,6 @@ def get_all_time_values(self) -> list[float]:
         """
         return list(self.data.keys())
 
-    @deprecated(
-        "`get_all_mesh_times()` is deprecated, use instead `get_all_time_values()`",
-        version="0.1.11",
-        removal="0.2.0",
-    )
-    def get_all_mesh_times(self) -> list[float]:
-        """DEPRECATED: Use :meth:`get_all_time_values` instead."""
-        return self.get_all_time_values()  # pragma: no cover
-
     def init_tree(self, time: Optional[float] = None) -> CGNSTree:
         """Initialize a CGNS tree structure at a specified time step or create a new one if it doesn't exist.
 

src/plaid/containers/sample.py

@@ -45,7 +45,6 @@
 )
 from plaid.utils import cgns_helper as CGH
 from plaid.utils.base import delegate_methods, safe_len
-from plaid.utils.deprecation import deprecated
 
 logger = logging.getLogger(__name__)
 
@@ -581,20 +580,6 @@ def extract_sample_from_identifier(
 
         return sample
 
-    @deprecated(
-        "`Dataset.from_features_identifier(...)` is deprecated, use instead `Dataset.extract_sample_from_identifier(...)`",
-        version="0.1.8",
-        removal="0.2",
-    )
-    def from_features_identifier(
-        self,
-        feature_identifiers: Union[FeatureIdentifier, list[FeatureIdentifier]],
-    ) -> Self:
-        """DEPRECATED: Use :meth:`Dataset.extract_sample_from_identifier` instead."""
-        return self.extract_sample_from_identifier(
-            feature_identifiers
-        )  # pragma: no cover
-
     def merge_features(self, sample: Self, in_place: bool = False) -> Self:
         """Merge features from another sample into the current sample.
 
@@ -639,18 +624,6 @@ def merge_features(self, sample: Self, in_place: bool = False) -> Self:
             in_place=in_place,
         )
 
-    # -------------------------------------------------------------------------#
-    @deprecated(
-        "`Sample.save(...)` is deprecated, use instead `Sample.save_to_dir(...)`",
-        version="0.1.8",
-        removal="0.2",
-    )
-    def save(
-        self, path: Union[str, Path], overwrite: bool = False, memory_safe: bool = False
-    ) -> None:
-        """DEPRECATED: use :meth:`Sample.save_to_dir` instead."""
-        self.save_to_dir(path, overwrite=overwrite, memory_safe=memory_safe)
-
     # -------------------------------------------------------------------------#
     def save_to_dir(
         self, path: Union[str, Path], overwrite: bool = False, memory_safe: bool = False
@@ -760,47 +733,6 @@ def load(self, path: Union[str, Path]) -> None:
 
                 (self.features.data[time],) = (tree,)
 
-        old_scalars_file = path / "scalars.csv"
-        if old_scalars_file.is_file():
-            self._load_old_scalars(old_scalars_file)
-
-        old_time_series_files = list(path.glob("time_series_*.csv"))
-        if len(old_time_series_files) > 0:
-            self._load_old_time_series(old_time_series_files)
-
-    @deprecated(
-        reason="This Sample was written with plaid<=0.1.9, save it with plaid>=0.1.10 to have all features embedded in the CGNS tree",
-        version="0.1.10",
-        removal="0.2.0",
-    )
-    def _load_old_scalars(self, scalars_file: Path):
-        names = np.loadtxt(scalars_file, dtype=str, max_rows=1, delimiter=",").reshape(
-            (-1,)
-        )
-        scalars = np.loadtxt(
-            scalars_file, dtype=float, skiprows=1, delimiter=","
-        ).reshape((-1,))
-        for name, value in zip(names, scalars):
-            self.add_scalar(name, value)
-
-    @deprecated(
-        reason="This Sample was written with plaid<=0.1.9, save it with plaid>=0.1.10 to have all features embedded in the CGNS tree",
-        version="0.1.10",
-        removal="0.2.0",
-    )
-    def _load_old_time_series(self, time_series_files: list[Path]):
-        for ts_fname in time_series_files:
-            names = np.loadtxt(ts_fname, dtype=str, max_rows=1, delimiter=",").reshape(
-                (-1,)
-            )
-            assert names[0] == "t"
-            times_and_val = np.loadtxt(ts_fname, dtype=float, skiprows=1, delimiter=",")
-            for i in range(times_and_val.shape[0]):
-                self.add_global(
-                    name=names[1],
-                    global_array=times_and_val[i, 1],
-                    time=times_and_val[i, 0],
-                )
 
     # # -------------------------------------------------------------------------#
     def __str__(self) -> str:

src/plaid/pipelines/sklearn_block_wrappers.py

@@ -249,7 +249,7 @@ def predict(self, dataset: Dataset) -> Dataset:
 
         dataset_predicted = Dataset.merge_dataset_by_features(
             [
-                dataset.from_tabular(
+                dataset.add_features_from_tabular(
                     y[
                         :,
                         None,
@@ -262,9 +262,7 @@ def predict(self, dataset: Dataset) -> Dataset:
                 for i_feat, feat_ids in enumerate(self.out_features_identifiers_)
             ]
         )
-        # dataset_predicted = dataset.add_features_from_tabular(
-        #     y, self.out_features_identifiers_, restrict_to_features=False
-        # )
+
         dataset_predicted = dataset.merge_features(dataset_predicted)
 
         return dataset_predicted

src/plaid/post/bisect.py

@@ -37,7 +37,7 @@ def prepare_datasets(
     assert ref_problem == pred_problem, "Reference and predicted dataset scalars differ"
 
     n_samples = len(ref_dataset)
-    out_scalars_names = problem_definition.get_output_scalars_names()
+    out_scalars_names = [f["name"] for f in problem_definition.get_out_features_identifiers()]
 
     ref_out_scalars = {}
     pred_out_scalars = {}