radnn 0.0.8__py3-none-any.whl → 0.1.0__py3-none-any.whl

radnn/models/cnn/cnn_stem_setup.py

@@ -0,0 +1,35 @@
+from enum import Enum
+
+class CNNSizeFactor(Enum):
+  TINY = 1
+  PETITE = 2
+  SMALL = 3
+  MEDIUM = 4
+
+
+class CNNStemSetup(object):
+  def __init__(self, size_factor: CNNSizeFactor, input_channels: int = 3, is_vit_like: bool = False):
+    self.size_factor: CNNSizeFactor = size_factor
+    self.input_channels = input_channels
+    self.windows = []
+    self.strides = []
+    
+    if is_vit_like:
+      self.windows = [4]
+      self.strides = [4]
+    else:
+      if self.size_factor is CNNSizeFactor.TINY:
+        self.windows = [3]
+        self.strides = [1]
+      elif self.size_factor is CNNSizeFactor.PETITE:
+        self.windows = [3]
+        self.strides = [2]
+      elif self.size_factor is CNNSizeFactor.SMALL:
+        self.windows = [5]
+        self.strides = [2]
+      elif self.size_factor is CNNSizeFactor.MEDIUM:
+        self.windows = [7, 3] #TODO: Check various papers
+        self.strides = [2, 2]
+      else:
+        raise ValueError(f"Unknown CNN scale factor: {self.size_factor.name}")
+    

radnn/models/model_factory.py

@@ -0,0 +1,85 @@
+# ======================================================================================
+#
+#     Rapid Deep Neural Networks
+#
+#     Licensed under the MIT License
+# ______________________________________________________________________________________
+# ......................................................................................
+
+# Copyright (c) 2018-2026 Pantelis I. Kaplanoglou
+
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+
+# The above copyright notice and this permission notice shall be included in all
+# copies or substantial portions of the Software.
+
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+
+# .......................................................................................
+from typing import Dict, Type, Any
+
+from radnn.models import ModelHyperparams, ModelInfo
+from radnn.errors import *
+
+
+# ======================================================================================================================
+class RadnnModel(type):
+  def __new__(mcls, name: str, bases: tuple[type, ...], ns: dict[str, Any]):
+    cls = super().__new__(mcls, name, bases, ns)
+    # Avoid registering the abstract base itself
+    if name != "ModelBuildAdapter":
+      key = ns.get("NAME", name)  # optional override via class attribute NAME
+      if key in ModelFactory.registry:
+        raise KeyError(f"The model '{key}' has been alrady registered ")
+      ModelFactory.registry[key] = cls
+    return cls
+
+
+# ======================================================================================================================
+class ModelBuildAdapter(metaclass=RadnnModel):
+  # --------------------------------------------------------------------------------------------------------------------
+  def __init__(self, **kwargs):
+    self.builder_kwargs = kwargs
+    self.hprm: dict | None = self.builder_kwargs.get("hprm", None)
+    self.hyperparams: ModelHyperparams | None = self.builder_kwargs.get("model_hyperparams", None)
+    if self.hyperparams is not None:
+      self.hprm = self.hyperparams.hprm
+    self.model = None
+    self.model_info: ModelInfo | None = None
+  
+  # --------------------------------------------------------------------------------------------------------------------
+  def build(self, **kwargs: Any):
+    raise NotImplementedError
+  # --------------------------------------------------------------------------------------------------------------------
+
+
+# ======================================================================================================================
+class ModelFactory(object):
+  registry: Dict[str, Type["ModelBuildAdapter"]] = {}
+  
+  @classmethod
+  def produce(cls, hyperparams: dict, **kwargs: Any) -> ModelBuildAdapter:
+    oModelBuilder = None
+    oModelHyperparams: ModelHyperparams = ModelHyperparams.from_dict(hyperparams)
+    assert oModelHyperparams.input_dims is not None, HPARAMS_DATA_INPUT_DIMS
+    assert len(oModelHyperparams.input_dims) > 0, HPARAMS_DATA_INPUT_DIMS
+    sModelAdapterKey = oModelHyperparams.base_name + "BuildAdapter"
+    if sModelAdapterKey in cls.registry:
+      cBuilder = ModelFactory.registry[sModelAdapterKey]
+      oModelBuilder = cBuilder(model_hyperparams=oModelHyperparams)
+      oModel = oModelBuilder.build(hprm=hyperparams, **kwargs)
+      oModelBuilder.model_info = ModelInfo(oModel, oModelHyperparams)
+    
+    return oModelBuilder
+# ======================================================================================================================

radnn/models/model_hyperparams.py

@@ -0,0 +1,128 @@
+# ======================================================================================
+#
+#     Rapid Deep Neural Networks
+#
+#     Licensed under the MIT License
+# ______________________________________________________________________________________
+# ......................................................................................
+
+# Copyright (c) 2018-2026 Pantelis I. Kaplanoglou
+
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+
+# The above copyright notice and this permission notice shall be included in all
+# copies or substantial portions of the Software.
+
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+
+# .......................................................................................
+import numpy as np
+from . import CNNSizeFactor
+
+
+class ModelHyperparams(object):
+  # --------------------------------------------------------------------------------------------------------------------
+  def __init__(self, hprm: dict | None = None):
+    self.hprm = hprm
+    self.model_name = None
+    self.base_name = ""
+    self.architecture = ""
+    self.variant = ""
+    self.class_count = 0
+    self.input_dims = []
+    self.dropout_prob = None
+    # self.kind = "" //TODO: Validate expected input according to model kind
+  
+  # --------------------------------------------------------------------------------------------------------------------
+  @property
+  def input_channels(self):
+    '''
+    :return: The channels from the dimensions of an image input sample
+    '''
+    return self.input_dims[-1]
+  
+  # --------------------------------------------------------------------------------------------------------------------
+  @property
+  def input_resolution(self):
+    '''
+    :return: The resolution from the dimensions of an image input sample
+    '''
+    return self.input_dims[:2]
+  
+  # --------------------------------------------------------------------------------------------------------------------
+  @property
+  def cnn_size_factor(self):
+    '''
+      Calculates the cnn size factor using the maximum dimension as log normalized where 0 is the size for MNIST input
+      and 1 the size for ImageNet input.
+    :return: 1  tiny images (MNIST, FMNIST, CIFAR10, CIFAR100)
+             2  small images (ImageNet 64x64)
+             3
+             4
+    '''
+    
+    def get_id(x):
+      if 0.0 <= x < 0.1:
+        return 1
+      elif 0.1 <= x < 0.4:
+        return 2
+      elif 0.4 <= x < 0.8:
+        return 3
+      elif 0.8 <= x <= 1.0:
+        return 4
+      else:
+        return None
+    
+    nHeight, nWidth = self.input_dims[:2]
+    nDim = nHeight
+    if nHeight < nWidth:
+      Dim = nWidth
+    nLogDim = np.log10(nDim)
+    nMin = np.log10(28)
+    nMax = np.log10(300)
+    nLogNormalized = (nLogDim - nMin) / (nMax - nMin)
+    nSizeFactor = CNNSizeFactor(get_id(nLogNormalized))
+    return nSizeFactor
+  
+  # --------------------------------------------------------------------------------------------------------------------
+  @classmethod
+  def from_dict(cls, hyperparams):
+    '''
+    Deserializes a hyperparameters object from a dictionary
+    
+    :param hyperparams: A dictionary containing the architectural hyperparameters for the model
+    :return: a new model hyperparams object
+    '''
+    oModelHyperparams = ModelHyperparams(hyperparams)
+    oModelHyperparams.model_name = hyperparams.get("Model.Name", "?")
+    
+    oModelHyperparams.base_name = oModelHyperparams.model_name
+    sParts = oModelHyperparams.base_name.split(".")
+    oModelHyperparams.architecture = "base"
+    if len(sParts) >= 2:
+      oModelHyperparams.base_name = sParts[0]
+      oModelHyperparams.architecture = sParts[1].lower()
+    
+    oModelHyperparams.variant = hyperparams.get("Model.Variant", "")
+    oModelHyperparams.name = hyperparams.get("Model.Name", "?")
+    oModelHyperparams.input_dims = hyperparams.get("Dataset.SampleDims", [])
+    oModelHyperparams.class_count = hyperparams.get("Dataset.ClassCount", 0)
+    nDropOutKeepProb = hyperparams.get("Training.Regularize.DropOut.KeepProb", None)
+    if nDropOutKeepProb is None:
+      oModelHyperparams.dropout_prob = hyperparams.get("Training.Regularize.DropOut.DropProb", 0.0)
+    else:
+      oModelHyperparams.dropout_prob = 1.0 - nDropOutKeepProb
+    
+    return oModelHyperparams
+  # --------------------------------------------------------------------------------------------------------------------

radnn/models/model_info.py

@@ -0,0 +1,91 @@
+# ======================================================================================
+#
+#     Rapid Deep Neural Networks
+#
+#     Licensed under the MIT License
+# ______________________________________________________________________________________
+# ......................................................................................
+
+# Copyright (c) 2018-2026 Pantelis I. Kaplanoglou
+
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+
+# The above copyright notice and this permission notice shall be included in all
+# copies or substantial portions of the Software.
+
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+
+# .......................................................................................
+from radnn import FileStore
+from radnn.data.structs import Tree, TreeNode
+
+from torchinfo import ModelStatistics, summary
+from torchinfo.layer_info import LayerInfo
+from .model_hyperparams import ModelHyperparams
+
+
+class ModelInfo(object):
+  def __init__(self, model=None, hyperparams: ModelHyperparams | dict = None, depth=6):
+    self.model = model
+    self.hyperparams = hyperparams
+    if isinstance(hyperparams, dict):
+      self.hyperparams = ModelHyperparams.from_dict(hyperparams)
+    self.model_tree: Tree = None
+    
+    if self.model is not None:
+      nInputDims = tuple([1, hyperparams.input_channels] + hyperparams.input_resolution)
+      self.summary: ModelStatistics = summary(model, nInputDims, device='cpu', depth=depth)
+      self.build_tree()
+  
+  # --------------------------------------------------------------------------------------------------------------------
+  @property
+  def filename(self):
+    return f'{self.hyperparams.model_name}_{str(self.hyperparams.input_dims).replace(", ", "x")[1:-1]}_cls{str(self.hyperparams.class_count)}'
+  
+  # --------------------------------------------------------------------------------------------------------------------
+  def save(self, fs: str | FileStore):
+    if isinstance(fs, str):
+      fs = FileStore(fs)
+    fs.artifact(self.filename + ".pkl").save(self.summary, is_overwriting=True)
+    with open(fs.file(self.filename + ".txt"), "w", encoding="utf-8") as oFile:
+      print(self.summary, file=oFile)
+    
+  # --------------------------------------------------------------------------------------------------------------------
+  def load(self, fs: str | FileStore):
+    bIsLoaded = False
+    if isinstance(fs, str):
+      fs = FileStore(fs)
+    self.summary = fs.artifact(self.filename + ".pkl").load()
+    bIsLoaded = self.summary is not None
+    if bIsLoaded:
+      self.build_tree()
+      '''
+      oNodeList = self.model_tree.traverse_depth_first()
+      for oNode in oNodeList:
+        print(oNode.path)
+      '''
+    return bIsLoaded
+  
+  # --------------------------------------------------------------------------------------------------------------------
+  def build_tree(self):
+    def recurse_build_subtree(source_node: LayerInfo, target_node: TreeNode):
+      for nIndex, oSourceChild in enumerate(source_node.children):
+        sNodeName = f"{oSourceChild.class_name}: {oSourceChild.depth}-{oSourceChild.depth_index}"
+        oTargetChild: TreeNode = target_node.new_child(sNodeName)
+        oTargetChild.value = oSourceChild
+        recurse_build_subtree(oSourceChild, oTargetChild)
+    
+    self.model_tree = Tree()
+    recurse_build_subtree(self.summary.summary_list[0], self.model_tree.root)
+  # --------------------------------------------------------------------------------------------------------------------

radnn/plots/plot_learning_curve.py

@@ -36,22 +36,33 @@ class PlotLearningCurve(object):
     self.model_name = model_name
     print("Keys in training process log:", self.metrics_dict.keys())
   # --------------------------------------------------------------------------------------
-  def prepare(self, metric_key="accuracy", custom_title=None, is_legend_right=False):
+  def prepare(self, metric_key="accuracy", custom_title=None, metric_low=0, metric_high=1.0, is_legend_right=False, is_keras=False):
+    sTrainMetricKey = metric_key
+    if not is_keras:
+      sTrainMetricKey = "train_" + metric_key
+    sValMetricKey = "val_" + metric_key
+    oLegend = ["training"]
+    bHasValidation = sValMetricKey in self.metrics_dict
+    oLegend.append("validation")
+    
+    
     plt.clf()
-    plt.plot(self.metrics_dict[metric_key])
-    sValidationMetricName = "val_" + metric_key
-    if sValidationMetricName in self.metrics_dict:
-      plt.plot(self.metrics_dict[sValidationMetricName])
+    plt.plot(self.metrics_dict[sTrainMetricKey])
+    if bHasValidation:
+      plt.plot(self.metrics_dict[sValMetricKey])
+      
     if custom_title is None:
       plt.title(self.model_name + ' ' + metric_key)
     else:
       plt.title(self.model_name + ' ' + custom_title)
-    plt.ylabel(metric_key)
+      
+    plt.ylabel(metric_key.capitalize().title())
     plt.xlabel("Epoch")
+    plt.ylim([metric_low, metric_high])
     if is_legend_right:
-      plt.legend(["train", 'validation'], loc="upper right")
+      plt.legend(oLegend, loc="upper right")
     else:
-      plt.legend(["train", "validation"], loc="upper left")
+      plt.legend(oLegend, loc="upper left")
     return self
   # --------------------------------------------------------------------------------------
   def prepare_cost(self, cost_function=None):

radnn/system/__init__.py

@@ -8,3 +8,4 @@ from .tee_logger import TeeLogger
 
 
 
+

radnn/system/files/__init__.py

@@ -4,4 +4,4 @@ from .picklefile import PickleFile
 from .textfile import TextFile
 from .csvfile import CSVFile
 from .filelist import FileList
-
+from .zipfile import ZipFile

radnn/system/files/csvfile.py

@@ -6,7 +6,7 @@
 # ______________________________________________________________________________________
 # ......................................................................................
 
-# Copyright (c) 2018-2025 Pantelis I. Kaplanoglou
+# Copyright (c) 2018-2026 Pantelis I. Kaplanoglou
 
 # Permission is hereby granted, free of charge, to any person obtaining a copy
 # of this software and associated documentation files (the "Software"), to deal
@@ -28,7 +28,7 @@
 
 # .......................................................................................
 import pandas as pd
-
+import numpy as np
 
 from .fileobject import FileObject
 
@@ -42,8 +42,40 @@ class CSVFile(FileObject):
     dResult = pd.read_csv(sFileName, delimiter=delimiter, encoding=encoding)
     return dResult
   # ----------------------------------------------------------------------------------
-  def save(self, obj, filename=None):
-    # TODO Save numpy array to pandas dataframe then to CSV
-    pass
+  # TODO: Iterate large csv files row by row as done in TextFile
+  #@property
+  #def rows(self):
+  # ----------------------------------------------------------------------------------
+  def save(self, obj, filename=None, headers=None, delimiter=",", encoding="utf-8"):
+    sFileName = self._useFileName(filename)
+
+    if isinstance(obj, dict):
+      nMaxRowCols = 0
+      oList = []
+      for k, v in obj.items():
+        oRow = [k]
+        if isinstance(v, list):
+          oRow.extend(v)
+        else:
+          oRow.append(v)
+        if len(oRow) > nMaxRowCols:
+          nMaxRowCols = len(oRow)
+        oList.append(oRow)
+
+      obj = []
+      for oRow in oList:
+        oRowToAdd = oRow + [None]*(nMaxRowCols - len(oRow))
+        obj.append(oRowToAdd)
+      if len(headers) < nMaxRowCols:
+        headers = headers + [None]*(nMaxRowCols - len(headers))
+
+    if isinstance(obj, list):
+      df = pd.DataFrame(obj, columns=headers)
+      df.to_csv(sFileName, index=False, header=(headers is not None), sep=delimiter, encoding=encoding)
 
+    elif isinstance(obj, np.ndarray):
+      if len(headers) < obj.shape[1]:
+        headers = headers + [None]*(obj.shape[1] - len(headers))
+      df = pd.DataFrame(obj, columns=headers)
+      df.to_csv(sFileName, index=False, header=(headers is not None), sep=delimiter, encoding=encoding)
   # ----------------------------------------------------------------------------------

radnn/system/files/filelist.py

@@ -1,3 +1,33 @@
+# ======================================================================================
+#
+#     Rapid Deep Neural Networks
+#
+#     Licensed under the MIT License
+# ______________________________________________________________________________________
+# ......................................................................................
+
+# Copyright (c) 2018-2026 Pantelis I. Kaplanoglou
+
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+
+# The above copyright notice and this permission notice shall be included in all
+# copies or substantial portions of the Software.
+
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+
+# .......................................................................................
+
 import os
 
 

radnn/system/files/fileobject.py

@@ -6,7 +6,7 @@
 # ______________________________________________________________________________________
 # ......................................................................................
 
-# Copyright (c) 2018-2025 Pantelis I. Kaplanoglou
+# Copyright (c) 2018-2026 Pantelis I. Kaplanoglou
 
 # Permission is hereby granted, free of charge, to any person obtaining a copy
 # of this software and associated documentation files (the "Software"), to deal
@@ -48,6 +48,16 @@ class FileObject(object):
       if not self.default_file_extension.startswith("."):
         self.default_file_extension = "." + self.default_file_extension
   # ----------------------------------------------------------------------------------
+  def close(self):
+    pass
+  # ----------------------------------------------------------------------------------
+  def __enter__(self):
+    return self
+  # ----------------------------------------------------------------------------------
+  def __exit__(self, exc_type, exc_value, traceback):
+    self.close()
+    return True
+  # ----------------------------------------------------------------------------------
   def _useFileName(self, filename):
     if filename is not None:
         self.filename = filename

radnn/system/files/imgfile.py

@@ -6,7 +6,7 @@
 # ______________________________________________________________________________________
 # ......................................................................................
 
-# Copyright (c) 2018-2025 Pantelis I. Kaplanoglou
+# Copyright (c) 2018-2026 Pantelis I. Kaplanoglou
 
 # Permission is hereby granted, free of charge, to any person obtaining a copy
 # of this software and associated documentation files (the "Software"), to deal

radnn/system/files/jsonfile.py

@@ -6,7 +6,7 @@
 # ______________________________________________________________________________________
 # ......................................................................................
 
-# Copyright (c) 2018-2025 Pantelis I. Kaplanoglou
+# Copyright (c) 2018-2026 Pantelis I. Kaplanoglou
 
 # Permission is hereby granted, free of charge, to any person obtaining a copy
 # of this software and associated documentation files (the "Software"), to deal
@@ -30,36 +30,64 @@
 import os
 import json
 import glob
+from collections import OrderedDict
+
 from .fileobject import FileObject
 
+#TODO: jsonpickle
+#https://stackoverflow.com/questions/3768895/how-to-make-a-class-json-serializable
+
 class JSONFile(FileObject):
   # ----------------------------------------------------------------------------------
   def __init__(self, filename, parent_folder=None, error_template=None):
     super(JSONFile, self).__init__(filename, parent_folder, error_template, "json")
   # ----------------------------------------------------------------------------------
-  def load(self, filename=None):
+  def load(self, filename=None, encoding=None):
     filename = self._useFileName(filename)
 
     dResult = None
     if os.path.exists(filename):
-      with open(filename) as oFile:
-        sJSON = oFile.read()
-      dResult = json.loads(sJSON)
+
+      sJSON = None
+      if os.path.isfile(filename):
+        oEncodingToTry = ["utf-8", "utf-16", "latin1", "ascii"]  # Add more if needed
+        if encoding is None:
+          bIsLoaded = False
+          for sEnc in oEncodingToTry:
+            try:
+              with open(filename, "r", encoding=sEnc) as oFile:
+                sJSON = oFile.read()
+              bIsLoaded = True
+              break
+            except (UnicodeDecodeError, UnicodeError):
+              continue
+          if not bIsLoaded:
+            raise ValueError("Unsupported encoding")
+        else:
+          with open(filename, "r", encoding=encoding) as oFile:
+            sJSON = oFile.read()
+      if sJSON is not None:
+        dResult = json.loads(sJSON, object_pairs_hook=OrderedDict)
     else:
       if self.error_template is not None:
         raise Exception(self.error_template % filename)
 
     return dResult
   # ----------------------------------------------------------------------------------
-  def save(self, obj, filename=None, is_sorted_keys=True):
+  def save(self, obj, filename=None, is_sorted_keys=False, is_utf8=False):
     filename = self._useFileName(filename)
 
     if obj is not None:
       if isinstance(obj, dict):
-        sJSON = json.dumps(obj, sort_keys=is_sorted_keys, indent=4)
+        sJSON = json.dumps(obj, sort_keys=is_sorted_keys, indent=4, ensure_ascii=is_utf8)
       else:
-        sJSON = json.dumps(obj, default=lambda o: obj.__dict__, sort_keys=is_sorted_keys, indent=4)
-      with open(filename, "w") as oFile:
+        sJSON = json.dumps(obj, default=lambda o: obj.__dict__, sort_keys=is_sorted_keys, indent=4, ensure_ascii=is_utf8)
+
+      sEncoding = "utf-8"
+      if not is_utf8:
+        sEncoding = "utf-16"
+
+      with open(filename, "w", encoding=sEncoding) as oFile:
         oFile.write(sJSON)
         oFile.close()
   # ----------------------------------------------------------------------------------

radnn/system/files/picklefile.py

@@ -6,7 +6,7 @@
 # ______________________________________________________________________________________
 # ......................................................................................
 
-# Copyright (c) 2018-2025 Pantelis I. Kaplanoglou
+# Copyright (c) 2018-2026 Pantelis I. Kaplanoglou
 
 # Permission is hereby granted, free of charge, to any person obtaining a copy
 # of this software and associated documentation files (the "Software"), to deal
@@ -64,7 +64,7 @@ class PickleFile(FileObject):
     del self._cached
     self._cached = None
   # ----------------------------------------------------------------------------------
-  def load(self, filename, is_python2_format=False, error_template=None):
+  def load(self, filename: str|None = None, is_python2_format: bool=False, error_template: str|None=None):
     """
     Deserializes the data from a pickle file if it exists.
     Parameters
@@ -101,7 +101,7 @@ class PickleFile(FileObject):
 
     return oData
   # ----------------------------------------------------------------------------------
-  def save(self, obj, filename=None, is_overwriting=False, extra_display_label=None):
+  def save(self, obj, filename: str | None =None, is_overwriting: bool =False, extra_display_label: str|None=None):
     """
     Serializes the data to a pickle file if it does not exists.
     Parameters