sinapsis-data-readers 0.1.12__tar.gz → 0.1.14__tar.gz

{sinapsis_data_readers-0.1.12/src/sinapsis_data_readers.egg-info → sinapsis_data_readers-0.1.14}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sinapsis-data-readers
-Version: 0.1.12
+Version: 0.1.14
 Summary: Templates to read data in different formats
 Author-email: SinapsisAI <dev@sinapsis.tech>
 Project-URL: Homepage, https://sinapsis.tech

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "sinapsis-data-readers"
-version = "0.1.12"
+version = "0.1.14"
 description = "Templates to read data in different formats"
 authors = [{ name = "SinapsisAI", email = "dev@sinapsis.tech" }]
 

sinapsis_data_readers-0.1.14/src/sinapsis_data_readers/helpers/csv_reader.py

@@ -0,0 +1,22 @@
+# -*- coding: utf-8 -*-
+import pandas as pd
+
+
+def read_file(file: str) -> pd.DataFrame:
+    """
+    Reads a CSV file and returns its contents as a pandas DataFrame.
+
+    Args:
+        file (str): The path to the CSV file to be read.
+
+    Returns:
+        pd.DataFrame: The data from the CSV file as a pandas DataFrame.
+
+    Raises:
+        ValueError: If the file does not have a .csv extension.
+    """
+    if not file.endswith("csv"):
+        raise ValueError("The file must have a .csv extension.")
+
+    data = pd.read_csv(file, header=0)
+    return data

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers/helpers/sklearn_dataset_subset.py

@@ -6,10 +6,22 @@ from sklearn import datasets
 _sklearn_supported_loaders = {
     name: getattr(datasets, name) for name in dir(datasets) if name.startswith(("load", "fetch"))
 }
+excluded_loaders = [
+    "fetch_lfw_pairs",
+    "fetch_20newsgroups",
+    "fetch_20newgroups_vectorized",
+    "load_sample_images",
+    "load_sample_image",
+    "load_svmlight_file",
+    "load_svmlight_files",
+    "fetch_rcv1",
+    "fetch_species_distribution",
+    "fetch_file",
+]
 
 
 def __getattr__(name: str) -> Callable:
-    if name in _sklearn_supported_loaders:
+    if name in _sklearn_supported_loaders and name not in excluded_loaders:
         return _sklearn_supported_loaders[name]
     raise AttributeError(f"Function `{name}` not found in sklearn.datasets.")
 

sinapsis_data_readers-0.1.14/src/sinapsis_data_readers/helpers/sktime_datasets_subset.py

@@ -0,0 +1,24 @@
+# -*- coding: utf-8 -*-
+""" Excluded sktime loaders"""
+from typing import Callable
+
+from sktime import datasets
+
+class_datasets = [
+"Airline",
+"Longley",
+"Lynx",
+"Macroeconomic",
+"ShampooSales",
+"Solar",
+"USChange"
+]
+
+
+def __getattr__(name: str) -> Callable:
+    if name in class_datasets:
+        return getattr(datasets, name)
+    raise AttributeError(f"Class `{name}` not found in sktime.datasets.")
+
+
+__all__ = class_datasets

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers/templates/__init__.py

@@ -10,6 +10,7 @@ _template_lookup = {
     "AudioReaderPydub": f"{_root_lib_path}.audio_readers.audio_reader_pydub",
     "AudioReaderSoundfile": f"{_root_lib_path}.audio_readers.audio_reader_soundfile",
     "AudioReaderToBytes": f"{_root_lib_path}.audio_readers.audio_reader_to_bytes",
+    "CSVDatasetReader": f"{_root_lib_path}.datasets_readers.csv_datasets",
     "CSVImageDataset": f"{_root_lib_path}.image_readers.csv_dataset_reader",
     "CocoDetectionDatasetCV2": f"{_root_lib_path}.image_readers.coco_dataset_reader",
     "CocoKeypointsDatasetCV2": f"{_root_lib_path}.image_readers.coco_dataset_reader",
@@ -19,7 +20,6 @@ _template_lookup = {
     "ExecuteNTimesLazyAudioReaderPydub": f"{_root_lib_path}.audio_readers.audio_reader_pydub",
     "ExecuteNTimesLazyAudioReaderSoundfile": f"{_root_lib_path}.audio_readers.audio_reader_soundfile",
     "FolderImageDatasetCV2": f"{_root_lib_path}.image_readers.image_folder_reader_cv2",
-    "FolderImageDatasetKornia": f"{_root_lib_path}.image_readers.image_folder_reader_kornia",
     "ImageDatasetSplitter": f"{_root_lib_path}.datasets_readers.dataset_splitter",
     "LazyAudioReaderPydub": f"{_root_lib_path}.audio_readers.audio_reader_pydub",
     "LazyAudioReaderSoundfile": f"{_root_lib_path}.audio_readers.audio_reader_soundfile",

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers/templates/audio_readers/audio_reader_pydub.py

@@ -134,6 +134,10 @@ class LazyAudioReaderPydub(AudioReaderPydub):
             from_bytes: False
     """
 
+    class AttributesBaseModel(AudioReaderPydub.AttributesBaseModel):
+        generic_key: str
+        audio_file_path: str | None = None  # type:ignore[assignment]
+
     def get_file_path_from_generic_data(self, container: DataContainer) -> None:
         """Method to retrieve the file path from the genetic data field of DataContainer.
         The method extracts the file path from the generic field and sets as attribute
@@ -141,8 +145,12 @@ class LazyAudioReaderPydub(AudioReaderPydub):
         Args:
             container (DataContainer): The DataContainer to extract the file path from
         """
-        file_path = container.generic_data.get("audio_path", "")
-        self.attributes.audio_file_path = file_path
+        if self.attributes.generic_key:
+            file_path = self._get_generic_data(container, self.attributes.generic_key)
+            if file_path:
+                self.attributes.audio_file_path = file_path
+            else:
+                self.logger.warning("No audio path in the existing container")
 
     def execute(self, container: DataContainer) -> DataContainer:
         self.get_file_path_from_generic_data(container)

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers/templates/audio_readers/audio_reader_soundfile.py

@@ -110,6 +110,10 @@ class LazyAudioReaderSoundfile(AudioReaderSoundfile):
             from_bytes: true
     """
 
+    class AttributesBaseModel(_AudioBaseReader.AttributesBaseModel):
+        generic_key: str
+        audio_file_path: str | None = None  # type:ignore[assignment]
+
     def get_file_path_from_generic_data(self, container: DataContainer) -> None:
         """Method to retrieve the file path from the genetic data field of DataContainer.
         The method extracts the file path from the generic field and sets as attribute
@@ -117,8 +121,12 @@ class LazyAudioReaderSoundfile(AudioReaderSoundfile):
         Args:
             container (DataContainer): The DataContainer to extract the file path from
         """
-        file_path = container.generic_data["audio_path"]
-        self.attributes.audio_file_path = file_path
+        if self.attributes.generic_key:
+            file_path = self._get_generic_data(container, self.attributes.generic_key)
+            if file_path:
+                self.attributes.audio_file_path = file_path
+            else:
+                self.logger.warning("No audio path in the existing container")
 
     def execute(self, container: DataContainer) -> DataContainer:
         self.get_file_path_from_generic_data(container)

sinapsis_data_readers-0.1.14/src/sinapsis_data_readers/templates/datasets_readers/csv_datasets.py

@@ -0,0 +1,26 @@
+# -*- coding: utf-8 -*-
+from sinapsis_core.data_containers.data_packet import DataContainer, TextPacket, TimeSeriesPacket
+from sinapsis_core.template_base.base_models import TemplateAttributes, TemplateAttributeType
+from sinapsis_core.template_base.template import Template
+
+from sinapsis_data_readers.helpers.csv_reader import read_file
+
+
+class CSVDatasetReader(Template):
+    class AttributesBaseModel(TemplateAttributes):
+        path_to_csv: str
+        store_as_time_series: bool = False
+        store_as_text_packet: bool = True
+
+    def __init__(self, attributes: TemplateAttributeType) -> None:
+        super().__init__(attributes)
+        self.csv_file = read_file(self.attributes.path_to_csv)
+
+    def execute(self, container: DataContainer) -> DataContainer:
+        if self.attributes.store_as_time_series:
+            packet = TimeSeriesPacket(content=self.csv_file)
+            container.time_series.append(packet)
+        if self.attributes.store_as_text_packet:
+            packet = TextPacket(content=self.csv_file)
+            container.texts.append(packet)
+        return container

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers/templates/datasets_readers/dataset_splitter.py

@@ -11,7 +11,7 @@ from sinapsis_core.template_base.base_models import TemplateAttributes
 from sklearn.model_selection import train_test_split
 
 ArrayDataFrameType = Union[list[np.ndarray], pd.DataFrame]
-StringDataFrameType = Union[list[str], pd.DataFrame]
+StringDataFrameType = Union[list[str | int], pd.DataFrame]
 OptionalArrayDataFrameType = Union[ArrayDataFrameType, None]
 
 OptionalStringDataFrameType = Union[StringDataFrameType, None]
@@ -30,9 +30,9 @@ class ImageDatasetSplit(BaseModel):
     """
 
     x_train: list[np.ndarray] = []
-    y_train: list[str] = []
+    y_train: list[str | int] = []
     x_test: list[np.ndarray] | None = None
-    y_test: list[str] | None = None
+    y_test: list[str | int] | None = None
 
     class Config:
         """allow arbitrary types"""
@@ -93,7 +93,11 @@ class DatasetSplitterBase(Template):
         x_train, x_test, y_train, y_test = x_data, None, y_data, None
         if self.attributes.train_size:
             x_train, x_test, y_train, y_test = train_test_split(
-                x_data, y_data, train_size=self.attributes.train_size, random_state=0
+                x_data,
+                y_data,
+                train_size=self.attributes.train_size,
+                test_size=1 - self.attributes.train_size,
+                random_state=0,
             )
         split_dataset = self.return_data_splitter_object(x_train=x_train, y_train=y_train, x_test=x_test, y_test=y_test)
         return split_dataset
@@ -126,7 +130,9 @@ class DatasetSplitterBase(Template):
         if not packet:
             self.logger.debug("No data to be processed by dataset splitter")
             return container
-
+        if len(packet) == 1:
+            self.logger.debug("Not enough entries to divide dataset, returning original container")
+            return container
         x_data, y_data = self.extract_x_y_from_packet(packet)
 
         custom_dataset = self.store_data_in_data_splitter(x_data, y_data)

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers/templates/datasets_readers/sklearn_datasets.py

@@ -1,5 +1,6 @@
 # -*- coding: utf-8 -*-
 
+import numpy as np
 import pandas as pd
 from sinapsis_core.data_containers.data_packet import DataContainer, TimeSeriesPacket
 from sinapsis_core.template_base import Template
@@ -14,6 +15,7 @@ from sinapsis_core.template_base.multi_execute_template import (
 )
 from sinapsis_core.utils.env_var_keys import SINAPSIS_BUILD_DOCS
 from sklearn.model_selection import train_test_split
+from sklearn.utils import Bunch
 
 from sinapsis_data_readers.helpers import sklearn_dataset_subset
 from sinapsis_data_readers.helpers.tags import Tags
@@ -65,7 +67,7 @@ class SKLearnDatasets(BaseDynamicWrapperTemplate):
         """
 
         split_dataset: bool = True
-        train_size: float = 1
+        train_size: float = 0.9
         store_as_time_series: bool = False
 
     def __init__(self, attributes: TemplateAttributeType) -> None:
@@ -73,7 +75,23 @@ class SKLearnDatasets(BaseDynamicWrapperTemplate):
         self.dataset_attributes = getattr(self.attributes, self.wrapped_callable.__name__)
 
     @staticmethod
-    def parse_results(results: pd.DataFrame) -> pd.DataFrame:
+    def process_bunch(bunch: Bunch) -> tuple:
+        data = bunch.get("data")
+        original_target = bunch.get("target")
+
+        target = np.asarray(original_target)
+        target = target.reshape(-1, 1) if target.ndim == 1 else target
+        feature_column = bunch.get("feature_names", None)
+        target_column = bunch.get("target_names", None)
+        if target.shape[1] == 1:
+            target_column = ["target"]
+        elif target_column is not None and len(target_column) == target.shape[1]:
+            target_column = list(target_column)
+        else:
+            target_column = [f"target_{i}" for i in range(target.shape[1])]
+        return data, target, feature_column, target_column
+
+    def parse_results(self, results: pd.DataFrame) -> tuple[pd.DataFrame, list, list, int]:
         """Parses the dataset as a pandas dataframe with the feature names as columns
 
         Args:
@@ -84,17 +102,51 @@ class SKLearnDatasets(BaseDynamicWrapperTemplate):
             the additional column for target values
 
         """
-
-        data_frame = pd.DataFrame(data=results.data, columns=results.feature_names)
-        data_frame[TARGET] = results.target
-        return data_frame
+        if isinstance(results, tuple):
+            data = results[0]
+            target = results[1]
+            feature_column = None
+            target_column = None
+        elif isinstance(results, Bunch):
+            data, target, feature_column, target_column = self.process_bunch(results)
+        else:
+            try:
+                data = results.data
+
+            except (KeyError, AttributeError, ValueError):
+                data = None
+            try:
+                target = results.target
+            except (KeyError, AttributeError, ValueError):
+                target = None
+            try:
+                feature_column = results.feature_names
+                target_column = results.target_names
+            except AttributeError:
+                feature_column = None
+                target_column = None
+        _, n_features = data.shape
+
+        feature_data_frame = pd.DataFrame(data=data, columns=feature_column)
+        target_data_frame = pd.DataFrame(data=target, columns=target_column)
+        data_frame = pd.concat([feature_data_frame, target_data_frame], axis=1)
+        return data_frame, feature_column, target_column, n_features
 
     @staticmethod
-    def split_dataset(results: pd.DataFrame, split_size: float) -> TabularDatasetSplit:
+    def split_dataset(
+        results: pd.DataFrame, feature_name_cols: list, target_name_cols: list, n_features: int, split_size: float
+    ) -> TabularDatasetSplit:
         """Method to split the dataset into training and testing samples"""
-        x_vals = results.drop(columns=[TARGET], axis=1)
-        y_vals = results[TARGET]
-        x_train, x_test, y_train, y_test = train_test_split(x_vals, y_vals, train_size=split_size, random_state=0)
+        if feature_name_cols:
+            X = results[feature_name_cols]
+            y = results[target_name_cols]
+        else:
+            X = results.iloc[:, :n_features]
+            y = results.iloc[:, n_features:]
+
+        # x_vals = results.drop(columns=[TARGET], axis=1)
+        # y_vals = results[TARGET]
+        x_train, x_test, y_train, y_test = train_test_split(X, y, train_size=split_size, random_state=0)
         split_data = TabularDatasetSplit(
             x_train=pd.DataFrame(x_train),
             x_test=pd.DataFrame(x_test),
@@ -106,13 +158,15 @@ class SKLearnDatasets(BaseDynamicWrapperTemplate):
 
     def execute(self, container: DataContainer) -> DataContainer:
         sklearn_dataset = self.wrapped_callable.__func__(**self.dataset_attributes.model_dump())
-        dataset = self.parse_results(sklearn_dataset)
+        dataset, feature_columns, target_columns, n_features = self.parse_results(sklearn_dataset)
         if self.attributes.store_as_time_series:
             time_series_packet = TimeSeriesPacket(content=dataset)
             container.time_series.append(time_series_packet)
 
         if self.attributes.split_dataset:
-            split_dataset = self.split_dataset(dataset, split_size=self.attributes.train_size)
+            split_dataset = self.split_dataset(
+                dataset, feature_columns, target_columns, n_features, split_size=self.attributes.train_size
+            )
             self._set_generic_data(container, split_dataset)
         if sklearn_dataset and not self.attributes.split_dataset:
             self._set_generic_data(container, dataset)

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers/templates/datasets_readers/sktime_datasets.py

@@ -1,4 +1,5 @@
 # -*- coding: utf-8 -*-
+from types import NoneType
 from typing import Any
 
 import pandas as pd
@@ -18,12 +19,24 @@ from sklearn.model_selection import train_test_split
 from sktime import datasets
 from sktime.split import temporal_train_test_split
 
+from sinapsis_data_readers.helpers import sktime_datasets_subset
+from sinapsis_data_readers.helpers.sktime_datasets_subset import class_datasets
 from sinapsis_data_readers.helpers.tags import Tags
 from sinapsis_data_readers.templates.datasets_readers.dataset_splitter import (
     TabularDatasetSplit,
 )
 
-EXCLUDE_MODULES = ["load_forecastingdata", "DATASET_NAMES_FPP3"]
+EXCLUDE_MODULES = ["load_forecastingdata", "DATASET_NAMES_FPP3", "BaseDataset",
+                   "load_gun_point_segmentation", "load_electric_devices_segments",
+                   "write_dataframe_to_tsfile",
+                   "write_ndarray_to_tsfile",
+                   "write_results_to_uea_format",
+                   "write_tabular_transformation_to_arff",
+                   "write_panel_to_tsfileWrapper",
+                   "_load_fpp3",
+                   "load_hierarchical_sales_toydata",
+                   "load_unitest_tsf"
+                   ] + class_datasets
 
 
 class SKTimeDatasets(BaseDynamicWrapperTemplate):
@@ -77,8 +90,10 @@ class SKTimeDatasets(BaseDynamicWrapperTemplate):
 
     def __init__(self, attributes: TemplateAttributeType) -> None:
         super().__init__(attributes)
-        self.dataset_attributes = getattr(self.attributes, self.wrapped_callable.__name__)
+        self.dataset_attributes = self.initialize_attributes()
 
+    def initialize_attributes(self):
+        return getattr(self.attributes, self.wrapped_callable.__name__)
     def split_time_series_dataset(self, dataset: Any) -> TabularDatasetSplit:
         """Split a time series dataset into training and testing sets
 
@@ -106,14 +121,22 @@ class SKTimeDatasets(BaseDynamicWrapperTemplate):
         Returns:
             TabularDatasetSplit: Object containing the split dataset.
         """
-        X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=self.attributes.train_size, random_state=0)
-        return TabularDatasetSplit(
-            x_train=pd.DataFrame(X_train),
-            x_test=pd.DataFrame(X_test),
-            y_train=pd.DataFrame(y_train),
-            y_test=pd.DataFrame(y_test),
-        )
-
+        try:
+            X_train, X_test, y_train, y_test = train_test_split(
+                X, y, train_size=self.attributes.train_size, random_state=0
+            )
+            return TabularDatasetSplit(
+                x_train=pd.DataFrame(X_train),
+                x_test=pd.DataFrame(X_test),
+                y_train=pd.DataFrame(y_train),
+                y_test=pd.DataFrame(y_test),
+            )
+        except ValueError:
+            self.logger.debug("Wrong format for split. original values")
+            return TabularDatasetSplit(x_train=pd.DataFrame(X), y_train=pd.DataFrame(y))
+
+    def create_dataset(self):
+        return self.wrapped_callable.__func__(**self.dataset_attributes.model_dump())
     def execute(self, container: DataContainer) -> DataContainer:
         """Execute the SKTimeDatasets template to load and process a dataset.
 
@@ -126,7 +149,7 @@ class SKTimeDatasets(BaseDynamicWrapperTemplate):
         Returns:
             DataContainer: The container with the dataset added to it.
         """
-        dataset = self.wrapped_callable.__func__(**self.dataset_attributes.model_dump())
+        dataset = self.create_dataset()
         split_dataset = dataset
         if isinstance(dataset, tuple):
             if self.attributes.split_dataset:
@@ -161,6 +184,36 @@ class ExecuteNTimesSKTimeDatasets(SKTimeDatasets):
     )
 
 
+class SKTimeClassDatasets(SKTimeDatasets):
+    WrapperEntry = WrapperEntryConfig(
+        wrapped_object=sktime_datasets_subset,
+        signature_from_doc_string=True,
+    )
+    def initialize_attributes(self):
+        return None
+    def create_dataset(self):
+        dataset = self.wrapped_callable.load("X", "y")
+        if isinstance(dataset[0], NoneType):
+            return dataset[1]
+        elif isinstance(dataset[1], NoneType):
+            return dataset[0]
+        return dataset
+
+@execute_template_n_times_wrapper
+class ExecuteNTimesSKTimeClassDatasets(SKTimeDatasets):
+    """This template extends the functionality of the SKTimeDatasets template
+    by loading the sktime dataset n times.
+
+    This is useful for running the same dataset loading operation multiple
+    times with different parameters or for benchmark purposes.
+    """
+
+    WrapperEntry = WrapperEntryConfig(
+        wrapped_object=sktime_datasets_subset,
+        signature_from_doc_string=True,
+        template_name_suffix="ExecuteNTimes",
+    )
+
 def __getattr__(name: str) -> Template:
     """
     Only create a template if it's imported, this avoids creating all the base models for all templates
@@ -170,10 +223,16 @@ def __getattr__(name: str) -> Template:
         return make_dynamic_template(name, SKTimeDatasets)
     if name in ExecuteNTimesSKTimeDatasets.WrapperEntry.module_att_names:
         return make_dynamic_template(name, ExecuteNTimesSKTimeDatasets)
+    if name in SKTimeClassDatasets.WrapperEntry.module_att_names:
+        return make_dynamic_template(name, SKTimeClassDatasets)
+    if name in ExecuteNTimesSKTimeClassDatasets.WrapperEntry.module_att_names:
+        return make_dynamic_template(name, ExecuteNTimesSKTimeClassDatasets)
     raise AttributeError(f"template `{name}` not found in {__name__}")
 
 
-__all__ = SKTimeDatasets.WrapperEntry.module_att_names + ExecuteNTimesSKTimeDatasets.WrapperEntry.module_att_names
+__all__ = (SKTimeDatasets.WrapperEntry.module_att_names + ExecuteNTimesSKTimeDatasets.WrapperEntry.module_att_names +
+           SKTimeClassDatasets.WrapperEntry.module_att_names +
+           ExecuteNTimesSKTimeClassDatasets.WrapperEntry.module_att_names)
 
 
 if SINAPSIS_BUILD_DOCS:

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers/templates/image_readers/coco_dataset_reader.py

@@ -54,10 +54,10 @@ class CocoImageDatasetBaseCV2(FolderImageDatasetCV2):
         annotations_path: str
 
     def __init__(self, attributes: TemplateAttributeType) -> None:
-        super().__init__(attributes)
-        self.annotations_file = os.path.join(self.attributes.data_dir, self.attributes.annotations_path)
+        self.annotations_file = os.path.join(attributes.get("data_dir"), attributes.get("annotations_path"))
         self.raw_annotations_dict: list[dict[str, dict[str, Any]]] = self.read_annotations_file(self.annotations_file)
         self.annotations = self.images_annotations()
+        super().__init__(attributes)
 
     @staticmethod
     def read_annotations_file(file: str) -> list[dict[str, dict[str, Any]]]:

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers/templates/image_readers/csv_dataset_reader.py

@@ -3,7 +3,6 @@
 from typing import cast
 
 import numpy as np
-import pandas as pd
 from sinapsis_core.data_containers.annotations import ImageAnnotations
 from sinapsis_core.data_containers.data_packet import ImagePacket
 from sinapsis_core.template_base.base_models import (
@@ -12,6 +11,7 @@ from sinapsis_core.template_base.base_models import (
     UIPropertiesMetadata,
 )
 
+from sinapsis_data_readers.helpers.csv_reader import read_file
 from sinapsis_data_readers.helpers.tags import Tags
 from sinapsis_data_readers.templates.base_file_data_loader import (
     ContentNotSetException,
@@ -19,26 +19,6 @@ from sinapsis_data_readers.templates.base_file_data_loader import (
 )
 
 
-def read_file(file: str) -> pd.DataFrame:
-    """
-    Reads a CSV file and returns its contents as a pandas DataFrame.
-
-    Args:
-        file (str): The path to the CSV file to be read.
-
-    Returns:
-        pd.DataFrame: The data from the CSV file as a pandas DataFrame.
-
-    Raises:
-        ValueError: If the file does not have a .csv extension.
-    """
-    if not file.endswith("csv"):
-        raise ValueError("The file must have a .csv extension.")
-
-    data = pd.read_csv(file, header=0)
-    return data
-
-
 class CSVImageDataset(_BaseDataReader):
     """
     A dataset reader for CSV-based image datasets, inheriting from _BaseDataReader.

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers/templates/video_readers/video_reader_dali.py

@@ -4,6 +4,7 @@
 from typing import Literal, cast
 
 import nvidia.dali.fn as fn
+import torch
 from nvidia.dali import pipeline_def
 from nvidia.dali.pipeline import DataNode, Pipeline
 from nvidia.dali.plugin.pytorch import DALIGenericIterator
@@ -149,6 +150,12 @@ class VideoReaderDali(BaseVideoReader):
             video_frames.append(self._make_image_packet(frame, frame_index=self.frame_count + idx))
         return video_frames
 
+    def reset_state(self, template_name: str | None = None) -> None:
+        _ = template_name
+        if self.attributes.device == "gpu":
+            torch.cuda.empty_cache()
+        super().reset_state(template_name)
+
 
 @multi_video_wrapper
 class MultiVideoReaderDali(VideoReaderDali):

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers/templates/video_readers/video_reader_ffmpeg.py

@@ -37,7 +37,6 @@ class VideoReaderFFMPEG(BaseVideoReader):
             video_file_path: '/path/to/video/file'
             batch_size: 1
             video_source: 4d2a355f-cda4-4742-9042-8e6ee842d1cf
-            device: cpu
             loop_forever: false
     """
 

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers/templates/video_readers/video_reader_torchcodec.py

@@ -1,4 +1,5 @@
 # -*- coding: utf-8 -*-
+import torch
 from sinapsis_core.data_containers.data_packet import ImagePacket
 from torchcodec.decoders import SimpleVideoDecoder
 
@@ -84,6 +85,12 @@ class VideoReaderTorchCodec(BaseVideoReader):
             video_frames.append(self._make_image_packet(frame, frame_index=self.frame_count + idx))
         return video_frames
 
+    def reset_state(self, template_name: str | None = None) -> None:
+        _ = template_name
+        if self.attributes.device == "gpu":
+            torch.cuda.empty_cache()
+        super().reset_state(template_name)
+
 
 @multi_video_wrapper
 class MultiVideoReaderTorchCodec(VideoReaderTorchCodec):
@@ -92,5 +99,4 @@ class MultiVideoReaderTorchCodec(VideoReaderTorchCodec):
     by adding as many video_readers as needed depending on the lenght of
     video_file_path list. It appends the dataframes of each of the videos to the
     ImagePacket object in DataContainer
-
     """

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14/src/sinapsis_data_readers.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sinapsis-data-readers
-Version: 0.1.12
+Version: 0.1.14
 Summary: Templates to read data in different formats
 Author-email: SinapsisAI <dev@sinapsis.tech>
 Project-URL: Homepage, https://sinapsis.tech

{sinapsis_data_readers-0.1.12 → sinapsis_data_readers-0.1.14}/src/sinapsis_data_readers.egg-info/SOURCES.txt

@@ -10,9 +10,11 @@ src/sinapsis_data_readers.egg-info/requires.txt
 src/sinapsis_data_readers.egg-info/top_level.txt
 src/sinapsis_data_readers/helpers/__init__.py
 src/sinapsis_data_readers/helpers/coco_dataclasses.py
+src/sinapsis_data_readers/helpers/csv_reader.py
 src/sinapsis_data_readers/helpers/file_path_helpers.py
 src/sinapsis_data_readers/helpers/image_color_space_converter.py
 src/sinapsis_data_readers/helpers/sklearn_dataset_subset.py
+src/sinapsis_data_readers/helpers/sktime_datasets_subset.py
 src/sinapsis_data_readers/helpers/tags.py
 src/sinapsis_data_readers/helpers/text_input_helpers.py
 src/sinapsis_data_readers/templates/__init__.py
@@ -23,6 +25,7 @@ src/sinapsis_data_readers/templates/audio_readers/audio_reader_soundfile.py
 src/sinapsis_data_readers/templates/audio_readers/audio_reader_to_bytes.py
 src/sinapsis_data_readers/templates/audio_readers/base_audio_reader.py
 src/sinapsis_data_readers/templates/datasets_readers/__init__.py
+src/sinapsis_data_readers/templates/datasets_readers/csv_datasets.py
 src/sinapsis_data_readers/templates/datasets_readers/dataset_splitter.py
 src/sinapsis_data_readers/templates/datasets_readers/sklearn_datasets.py
 src/sinapsis_data_readers/templates/datasets_readers/sktime_datasets.py