nnodely 1.3.0__tar.gz → 1.3.1__tar.gz

{nnodely-1.3.0/nnodely.egg-info → nnodely-1.3.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: nnodely
-Version: 1.3.0
+Version: 1.3.1
 Summary: Model-structured neural network framework for the modeling and control of physical systems
 Author-email: Gastone Pietro Rosati Papini <tonegas@gmail.com>
 License: MIT License
@@ -51,7 +51,7 @@ Requires-Dist: onnxruntime
 [![Coverage Status](https://readthedocs.org/projects/nnodely/badge/?version=latest&style=default)](https://nnodely.readthedocs.io/en/latest/)
 
 <a name="readme-top"></a>
-## Model-structured neural network framework for the modeling and control of physical systems
+# Model-structured neural network framework for the modeling and control of physical systems
 
 Modeling, control, and estimation of physical systems are fundamental to a wide range of engineering applications. 
 However, integrating data-driven methods like neural networks into these domains presents significant challenges, 
@@ -62,7 +62,7 @@ MSNNs merge the flexibility of neural networks with structures grounded in physi
 providing a powerful tool for representing and managing complex physical systems. Moreover, a MSNN thanks to the reduced
 number of parameters needs fewer data for training and can be used in real-time applications.
 
-### Why use nnodely
+## Why use nnodely
 
 The framework's goal is to allow the users fast modeling and control of a any mechanical systems using 
 a hybrind approach between neural networks and physical models.
@@ -92,13 +92,13 @@ a Pytorch (nnodely independent) and ONNX export.
       <a href="#settingstarted">Getting Started</a>
     </li>
     <li>
-      <a href="#basicfunctionalities">Basic Functionalities</a>
+      <a href="#basicfunctionalities">Basic Example</a>
       <ul>
-        <li><a href="#structuredneuralmodel">Build the structured neural model</a></li>
-        <li><a href="#neuralizemodel">Neuralize the structured neural model</a></li>
+        <li><a href="#structuredneuralmodel">Build the neural model</a></li>
+        <li><a href="#neuralizemodel">Neuralize the neural model</a></li>
         <li><a href="#loaddataset">Load the dataset</a></li>
-        <li><a href="#trainmodel">Train the structured neural network</a></li>
-        <li><a href="#testmodel">Test the structured neural model</a></li>
+        <li><a href="#trainmodel">Train the neural model</a></li>
+        <li><a href="#testmodel">Test the neural model</a></li>
       </ul>
     </li>
     <li>
@@ -109,15 +109,14 @@ a Pytorch (nnodely independent) and ONNX export.
         <li><a href="#examplesfolder">examples folder</a></li>
       </ul>
     </li>
-    <li>
-      <a href="#license">License</a>
-    </li>
+    <li><a href="#license">License</a></li>
+    <li><a href="#references">References</a></li>
   </ol>
 </details>
 
 <!-- GETTING STARTED -->
 <a name="settingstarted"></a>
-### Getting Started
+## Getting Started
 You can install the nnodely framework from PyPI via:
   ```sh
   pip install nnodely
@@ -191,7 +190,7 @@ Our formulation is more general and can take into account the noise of the measu
 The estimator can also be seen as the composition of the force contributions due to the position and velocity of the mass plus the contribution of external forces.
 
 <a name="neuralizemodel"></a>
-### Neuralize the structured neural model
+### Neuralize the neural model
 Let's now try to train our observer using the data we have.
 We perform:
 ```python
@@ -233,7 +232,7 @@ mass_spring_damper.loadData(name='mass_spring_dataset', source=data_folder, form
 Finally, the dataset is loaded. **nnodely** loads all the files that are in a source folder.
 
 <a name="trainmodel"></a>
-### Train the structured neural network
+### Train the neural network
 Using the dataset created the training is performed on the model.
 
 ```python
@@ -241,7 +240,7 @@ mass_spring_damper.trainModel()
 ```
 
 <a name="testmodel"></a>
-### Test the structured neural model
+### Test the neural model
 In order to test the results we need to create a input, in this case is defined by:
 1. `x` with 5 sample because the sample time is 0.2 and the window of `x`is 1 second.
 2. `F` is one sample because only the last sample is needed.
@@ -329,6 +328,7 @@ This project is released under the license [License: MIT](https://opensource.org
 
 <p align="right">(<a href="#readme-top">back to top</a>)</p>
 
+<a name="references"></a>
 ## References
 <a id="1">[1]</a> 
 Mauro Da Lio, Daniele Bortoluzzi, Gastone Pietro Rosati Papini. (2019). 
@@ -360,3 +360,5 @@ and [[code extended]](https://github.com/tonegas/nnodely-applications/blob/main/
 Hector Perez-Villeda, Justus Piater, Matteo Saveriano. (2023).
 Learning and extrapolation of robotic skills using task-parameterized equation learner networks.
 Robotics and Autonomous Systems. https://doi.org/10.1016/j.robot.2022.104309
+
+<p align="right">(<a href="#readme-top">back to top</a>)</p>

{nnodely-1.3.0 → nnodely-1.3.1}/README.md

@@ -7,7 +7,7 @@
 [![Coverage Status](https://readthedocs.org/projects/nnodely/badge/?version=latest&style=default)](https://nnodely.readthedocs.io/en/latest/)
 
 <a name="readme-top"></a>
-## Model-structured neural network framework for the modeling and control of physical systems
+# Model-structured neural network framework for the modeling and control of physical systems
 
 Modeling, control, and estimation of physical systems are fundamental to a wide range of engineering applications. 
 However, integrating data-driven methods like neural networks into these domains presents significant challenges, 
@@ -18,7 +18,7 @@ MSNNs merge the flexibility of neural networks with structures grounded in physi
 providing a powerful tool for representing and managing complex physical systems. Moreover, a MSNN thanks to the reduced
 number of parameters needs fewer data for training and can be used in real-time applications.
 
-### Why use nnodely
+## Why use nnodely
 
 The framework's goal is to allow the users fast modeling and control of a any mechanical systems using 
 a hybrind approach between neural networks and physical models.
@@ -48,13 +48,13 @@ a Pytorch (nnodely independent) and ONNX export.
       <a href="#settingstarted">Getting Started</a>
     </li>
     <li>
-      <a href="#basicfunctionalities">Basic Functionalities</a>
+      <a href="#basicfunctionalities">Basic Example</a>
       <ul>
-        <li><a href="#structuredneuralmodel">Build the structured neural model</a></li>
-        <li><a href="#neuralizemodel">Neuralize the structured neural model</a></li>
+        <li><a href="#structuredneuralmodel">Build the neural model</a></li>
+        <li><a href="#neuralizemodel">Neuralize the neural model</a></li>
         <li><a href="#loaddataset">Load the dataset</a></li>
-        <li><a href="#trainmodel">Train the structured neural network</a></li>
-        <li><a href="#testmodel">Test the structured neural model</a></li>
+        <li><a href="#trainmodel">Train the neural model</a></li>
+        <li><a href="#testmodel">Test the neural model</a></li>
       </ul>
     </li>
     <li>
@@ -65,15 +65,14 @@ a Pytorch (nnodely independent) and ONNX export.
         <li><a href="#examplesfolder">examples folder</a></li>
       </ul>
     </li>
-    <li>
-      <a href="#license">License</a>
-    </li>
+    <li><a href="#license">License</a></li>
+    <li><a href="#references">References</a></li>
   </ol>
 </details>
 
 <!-- GETTING STARTED -->
 <a name="settingstarted"></a>
-### Getting Started
+## Getting Started
 You can install the nnodely framework from PyPI via:
   ```sh
   pip install nnodely
@@ -147,7 +146,7 @@ Our formulation is more general and can take into account the noise of the measu
 The estimator can also be seen as the composition of the force contributions due to the position and velocity of the mass plus the contribution of external forces.
 
 <a name="neuralizemodel"></a>
-### Neuralize the structured neural model
+### Neuralize the neural model
 Let's now try to train our observer using the data we have.
 We perform:
 ```python
@@ -189,7 +188,7 @@ mass_spring_damper.loadData(name='mass_spring_dataset', source=data_folder, form
 Finally, the dataset is loaded. **nnodely** loads all the files that are in a source folder.
 
 <a name="trainmodel"></a>
-### Train the structured neural network
+### Train the neural network
 Using the dataset created the training is performed on the model.
 
 ```python
@@ -197,7 +196,7 @@ mass_spring_damper.trainModel()
 ```
 
 <a name="testmodel"></a>
-### Test the structured neural model
+### Test the neural model
 In order to test the results we need to create a input, in this case is defined by:
 1. `x` with 5 sample because the sample time is 0.2 and the window of `x`is 1 second.
 2. `F` is one sample because only the last sample is needed.
@@ -285,6 +284,7 @@ This project is released under the license [License: MIT](https://opensource.org
 
 <p align="right">(<a href="#readme-top">back to top</a>)</p>
 
+<a name="references"></a>
 ## References
 <a id="1">[1]</a> 
 Mauro Da Lio, Daniele Bortoluzzi, Gastone Pietro Rosati Papini. (2019). 
@@ -315,4 +315,6 @@ and [[code extended]](https://github.com/tonegas/nnodely-applications/blob/main/
 <a id="6">[6]</a> 
 Hector Perez-Villeda, Justus Piater, Matteo Saveriano. (2023).
 Learning and extrapolation of robotic skills using task-parameterized equation learner networks.
-Robotics and Autonomous Systems. https://doi.org/10.1016/j.robot.2022.104309
+Robotics and Autonomous Systems. https://doi.org/10.1016/j.robot.2022.104309
+
+<p align="right">(<a href="#readme-top">back to top</a>)</p>

{nnodely-1.3.0 → nnodely-1.3.1}/nnodely/__init__.py

@@ -1,5 +1,5 @@
 
-__version__ = '1.3.0'
+__version__ = '1.3.1'
 
 import sys
 major, minor = sys.version_info.major, sys.version_info.minor

{nnodely-1.3.0 → nnodely-1.3.1}/nnodely/activation.py

@@ -85,17 +85,17 @@ class Identity(Stream, ToStream):
 
 class Softmax(Stream, ToStream):
     """
-        Implement the Softmax relation function.
+    Implement the Softmax relation function.
 
-        See also:
-            Official PyTorch Softmax documentation: 
-            `torch.nn.Softmax <https://pytorch.org/docs/stable/generated/torch.nn.Softmax.html>`_
+    See also:
+        Official PyTorch Softmax documentation:
+        `torch.nn.Softmax <https://pytorch.org/docs/stable/generated/torch.nn.Softmax.html>`_
 
-        :param obj: The relation stream.
-        :type obj: Stream 
+    :param obj: The relation stream.
+    :type obj: Stream
 
-        Example:
-            >>> x = Softmax(x)
+    Example:
+        >>> x = Softmax(x)
     """
     @enforce_types
     def __init__(self, obj:Stream|Parameter|Constant|float|int) -> Stream:
@@ -106,22 +106,23 @@ class Softmax(Stream, ToStream):
         self.json['Relations'][self.name] = [softmax_relation_name, [obj.name]]
 
 class Sigmoid(Stream, ToStream):
-    """
-        Implement the Sigmoid relation function.
-        The Sigmoid function is defined as:
+    r"""
+    Implement the Sigmoid relation function.
+    The Sigmoid function is defined as:
 
-        see also:
-            Official PyTorch Softmax documentation:
-            `Sigmoid function <https://pytorch.org/docs/stable/generated/torch.nn.Sigmoid.html#torch.nn.Sigmoid>`_
+    see also:
+        Official PyTorch Softmax documentation:
+        `Sigmoid function <https://pytorch.org/docs/stable/generated/torch.nn.Sigmoid.html#torch.nn.Sigmoid>`_
 
-        :math:
-            \sigma(x) = \frac{1}{1 + e^{-x}}
+    .. math::
 
-        :param obj: The relation stream.
-        :type obj: Stream 
+        \sigma(x) = \frac{1}{1 + e^{-x}}
 
-        Example:
-            >>> x = Sigmoid(x)
+    :param obj: The relation stream.
+    :type obj: Stream
+
+    Example:
+        >>> x = Sigmoid(x)
     """
     @enforce_types
     def __init__(self, obj:Stream|Parameter|Constant|float|int) -> Stream:

{nnodely-1.3.0 → nnodely-1.3.1}/nnodely/arithmetic.py

@@ -131,8 +131,8 @@ class Pow(Stream, ToStream):
         (it is also possible to use the classical math operator '**')
 
         See also:
-            Official PyTorch Add documentation: 
-            `torch.add <https://pytorch.org/docs/stable/generated/torch.pow.html>`_
+            Official PyTorch pow documentation:
+            `torch.pow <https://pytorch.org/docs/stable/generated/torch.pow.html>`_
 
         :param input: the base of the power function
         :type obj: Tensor

{nnodely-1.3.0 → nnodely-1.3.1}/nnodely/initializer.py

@@ -7,8 +7,8 @@ def init_constant(indexes, params_size, dict_param = {'value':1}):
     ----------
     dict_param : dict, optional
         Dictionary containing the initialization parameters. Default is {'value': 1}.
-        - value : int or float
-            The constant value to initialize the parameters with.
+            value : int or float
+                The constant value to initialize the parameters with.
     """
     return dict_param['value']
 
@@ -24,12 +24,12 @@ def init_negexp(indexes, params_size, dict_param = {'size_index':0, 'first_value
         List of sizes for each dimension of the parameters.
     dict_param : dict, optional
         Dictionary containing the initialization parameters. Default is {'size_index': 0, 'first_value': 1, 'lambda': 3}.
-        - size_index : int
-            The index of the dimension to apply the exponential function.
-        - first_value : int or float
-            The value at the start of the range.
-        - lambda : int or float
-            The decay rate parameter of the exponential function.
+            size_index : int
+                The index of the dimension to apply the exponential function.
+            first_value : int or float
+                The value at the start of the range.
+            lambda : int or float
+                The decay rate parameter of the exponential function.
     """
     import numpy as np
     size_index = dict_param['size_index']
@@ -49,14 +49,14 @@ def init_exp(indexes, params_size, dict_param = {'size_index':0, 'max_value':1,
         List of sizes for each dimension of the parameters.
     dict_param : dict, optional
         Dictionary containing the initialization parameters. Default is {'size_index': 0, 'max_value': 1, 'lambda': 3, 'monotonicity': 'decreasing'}.
-        - size_index : int
-            The index of the dimension to apply the exponential function.
-        - max_value : int or float
-            The maximum value of the exponential function.
-        - lambda : int or float
-            The rate parameter of the exponential function.
-        - monotonicity : str
-            The monotonicity of the exponential function. Can be 'increasing' or 'decreasing'.
+            size_index : int
+                The index of the dimension to apply the exponential function.
+            max_value : int or float
+                The maximum value of the exponential function.
+            lambda : int or float
+                The rate parameter of the exponential function.
+            monotonicity : str
+                The monotonicity of the exponential function. Can be 'increasing' or 'decreasing'.
 
     Raises
     ------
@@ -90,12 +90,12 @@ def init_lin(indexes, params_size, dict_param = {'size_index':0, 'first_value':1
         List of sizes for each dimension of the parameters.
     dict_param : dict, optional
         Dictionary containing the initialization parameters. Default is {'size_index': 0, 'first_value': 1, 'last_value': 0}.
-        - size_index : int
-            The index of the dimension to apply the linear function.
-        - first_value : int or float
-            The value at the start of the range.
-        - last_value : int or float
-            The value at the end of the range.
+            size_index : int
+                The index of the dimension to apply the linear function.
+            first_value : int or float
+                The value at the start of the range.
+            last_value : int or float
+                The value at the end of the range.
     """
     size_index = dict_param['size_index']
     x = 0 if params_size[size_index]-1 == 0 else indexes[size_index]/(params_size[size_index]-1)

{nnodely-1.3.0 → nnodely-1.3.1}/nnodely/input.py

@@ -112,19 +112,17 @@ class InputState(NeuObj):
 
         Examples
         --------
-        Select a sample window considering a signal T = [-3,-2,-1,0,1,2] where the time vector 0 represent the last passed instant
-        If sw is an integer #1 represent the number of step in the past
-            T.s(2) = [-1, 0]    # represents two sample step in the past
+        Select a sample window considering a signal T = [-3,-2,-1,0,1,2] where the time vector 0 represent the last passed instant. If sw is an integer #1 represent the number of step in the past
+            >>> T.s(2) #= [-1, 0] represents two sample step in the past
 
         If sw is a list [#1,#2] the numbers represent the sample indexes in the vector with the second element excluded
-            T.s([-2,0])  = [-1, 0]    # represents two time step in the past zero in the future
-            T.s([0,1])   = [1]        # the first time in the future
-            T.s([-4,-2]) = [-3,-2]
-
-        The total number of samples can be computed #2-#1
-        The offset represent the index of the vector that need to be used to offset the window
-            T.s(2,offset=-2)       = [0, 1]      # the value of the window is [-1,0]
-            T.s([-2,2],offset=-1)  = [-1,0,1,2]  # the value of the window is [-1,0,1,2]
+            >>> T.s([-2,0])  #= [-1, 0] represents two time step in the past zero in the future
+            >>> T.s([0,1])   #= [1]     the first time in the future
+            >>> T.s([-4,-2]) #= [-3,-2]
+
+        The total number of samples can be computed #2-#1. The offset represent the index of the vector that need to be used to offset the window
+            >>> T.s(2,offset=-2)       #= [0, 1]      the value of the window is [-1,0]
+            >>> T.s([-2,2],offset=-1)  #= [-1,0,1,2]  the value of the window is [-1,0,1,2]
         """
         dim = copy.deepcopy(self.dim)
         json = copy.deepcopy(self.json)

{nnodely-1.3.0 → nnodely-1.3.1}/nnodely/nnodely.py

@@ -1875,9 +1875,8 @@ class Modely:
         """
         Exports the neural network model to an ONNX file.
 
-        -----
         .. note::
-            The input_order must contain all the inputs and states of the model in the order that you want to export them.
+            The input_order may contain all the inputs and states of the model in the order that you want to export them.
 
         Parameters
         ----------
@@ -1932,7 +1931,6 @@ class Modely:
         """
         Run an inference session using an onnx model previously exported using the nnodely framework. 
 
-        -----
         .. note:: Feed-Forward ONNX model
             For feed-forward models, the onnx model expect all the inputs and states to have 3 dimensions. The first dimension is the batch size, the second is the time window and the third is the feature dimension.
         .. note:: Recurrent ONNX model

{nnodely-1.3.0 → nnodely-1.3.1}/nnodely/relation.py

@@ -109,6 +109,9 @@ class Relation():
         return Neg(self)
 
 class Stream(Relation):
+    """
+    Represents a stream of data inside the neural network. A Stream is automatically create when you operate over a Input, State, Parameter, or Constant object.
+    """
     count = 0
     @classmethod
     def resetCount(self):

{nnodely-1.3.0 → nnodely-1.3.1/nnodely.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: nnodely
-Version: 1.3.0
+Version: 1.3.1
 Summary: Model-structured neural network framework for the modeling and control of physical systems
 Author-email: Gastone Pietro Rosati Papini <tonegas@gmail.com>
 License: MIT License
@@ -51,7 +51,7 @@ Requires-Dist: onnxruntime
 [![Coverage Status](https://readthedocs.org/projects/nnodely/badge/?version=latest&style=default)](https://nnodely.readthedocs.io/en/latest/)
 
 <a name="readme-top"></a>
-## Model-structured neural network framework for the modeling and control of physical systems
+# Model-structured neural network framework for the modeling and control of physical systems
 
 Modeling, control, and estimation of physical systems are fundamental to a wide range of engineering applications. 
 However, integrating data-driven methods like neural networks into these domains presents significant challenges, 
@@ -62,7 +62,7 @@ MSNNs merge the flexibility of neural networks with structures grounded in physi
 providing a powerful tool for representing and managing complex physical systems. Moreover, a MSNN thanks to the reduced
 number of parameters needs fewer data for training and can be used in real-time applications.
 
-### Why use nnodely
+## Why use nnodely
 
 The framework's goal is to allow the users fast modeling and control of a any mechanical systems using 
 a hybrind approach between neural networks and physical models.
@@ -92,13 +92,13 @@ a Pytorch (nnodely independent) and ONNX export.
       <a href="#settingstarted">Getting Started</a>
     </li>
     <li>
-      <a href="#basicfunctionalities">Basic Functionalities</a>
+      <a href="#basicfunctionalities">Basic Example</a>
       <ul>
-        <li><a href="#structuredneuralmodel">Build the structured neural model</a></li>
-        <li><a href="#neuralizemodel">Neuralize the structured neural model</a></li>
+        <li><a href="#structuredneuralmodel">Build the neural model</a></li>
+        <li><a href="#neuralizemodel">Neuralize the neural model</a></li>
         <li><a href="#loaddataset">Load the dataset</a></li>
-        <li><a href="#trainmodel">Train the structured neural network</a></li>
-        <li><a href="#testmodel">Test the structured neural model</a></li>
+        <li><a href="#trainmodel">Train the neural model</a></li>
+        <li><a href="#testmodel">Test the neural model</a></li>
       </ul>
     </li>
     <li>
@@ -109,15 +109,14 @@ a Pytorch (nnodely independent) and ONNX export.
         <li><a href="#examplesfolder">examples folder</a></li>
       </ul>
     </li>
-    <li>
-      <a href="#license">License</a>
-    </li>
+    <li><a href="#license">License</a></li>
+    <li><a href="#references">References</a></li>
   </ol>
 </details>
 
 <!-- GETTING STARTED -->
 <a name="settingstarted"></a>
-### Getting Started
+## Getting Started
 You can install the nnodely framework from PyPI via:
   ```sh
   pip install nnodely
@@ -191,7 +190,7 @@ Our formulation is more general and can take into account the noise of the measu
 The estimator can also be seen as the composition of the force contributions due to the position and velocity of the mass plus the contribution of external forces.
 
 <a name="neuralizemodel"></a>
-### Neuralize the structured neural model
+### Neuralize the neural model
 Let's now try to train our observer using the data we have.
 We perform:
 ```python
@@ -233,7 +232,7 @@ mass_spring_damper.loadData(name='mass_spring_dataset', source=data_folder, form
 Finally, the dataset is loaded. **nnodely** loads all the files that are in a source folder.
 
 <a name="trainmodel"></a>
-### Train the structured neural network
+### Train the neural network
 Using the dataset created the training is performed on the model.
 
 ```python
@@ -241,7 +240,7 @@ mass_spring_damper.trainModel()
 ```
 
 <a name="testmodel"></a>
-### Test the structured neural model
+### Test the neural model
 In order to test the results we need to create a input, in this case is defined by:
 1. `x` with 5 sample because the sample time is 0.2 and the window of `x`is 1 second.
 2. `F` is one sample because only the last sample is needed.
@@ -329,6 +328,7 @@ This project is released under the license [License: MIT](https://opensource.org
 
 <p align="right">(<a href="#readme-top">back to top</a>)</p>
 
+<a name="references"></a>
 ## References
 <a id="1">[1]</a> 
 Mauro Da Lio, Daniele Bortoluzzi, Gastone Pietro Rosati Papini. (2019). 
@@ -360,3 +360,5 @@ and [[code extended]](https://github.com/tonegas/nnodely-applications/blob/main/
 Hector Perez-Villeda, Justus Piater, Matteo Saveriano. (2023).
 Learning and extrapolation of robotic skills using task-parameterized equation learner networks.
 Robotics and Autonomous Systems. https://doi.org/10.1016/j.robot.2022.104309
+
+<p align="right">(<a href="#readme-top">back to top</a>)</p>

{nnodely-1.3.0 → nnodely-1.3.1}/nnodely.egg-info/SOURCES.txt

@@ -50,6 +50,7 @@ nnodely/visualizer/dynamicmpl/fuzzyplot.py
 nnodely/visualizer/dynamicmpl/resultsplot.py
 nnodely/visualizer/dynamicmpl/trainingplot.py
 tests/test_dataset.py
+tests/test_documentation.py
 tests/test_export.py
 tests/test_export_recurrent.py
 tests/test_input_dimensions.py

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_dataset.py

@@ -598,8 +598,4 @@ class ModelyCreateDatasetTest(unittest.TestCase):
         test.loadData(name='dataset', source=data_folder, format=data_struct, skiplines=1, delimiter=' ')
         self.assertListEqual(list(test.data['dataset']['x'].shape), [42, 6, 1])
         self.assertListEqual(list(test.data['dataset']['y'].shape), [42, 4, 1])
-        self.assertListEqual(test.multifile['dataset'], [4, 18, 42])
-
-
-if __name__ == '__main__':
-    unittest.main()
+        self.assertListEqual(test.multifile['dataset'], [4, 18, 42])

nnodely-1.3.1/tests/test_documentation.py

@@ -0,0 +1,25 @@
+import unittest
+import subprocess
+import os
+
+class TestDocumentation(unittest.TestCase):
+    def test_generate_docs(self):
+        # Path to the Sphinx documentation source directory
+        docs_source_dir = os.path.join(os.path.dirname(__file__), '..', 'docs')
+
+        # Path to the output directory for the generated documentation
+        docs_output_dir = os.path.join(docs_source_dir, '_build', 'html')
+
+        # Command to generate the documentation
+        command = ['sphinx-build', '--fail-on-warning', '-b',  'html', docs_source_dir, docs_output_dir]
+
+        # Run the command and capture the output
+        result = subprocess.run(command, capture_output=True, text=True)
+
+        # Check if the command was successful
+        self.assertEqual(result.returncode, 0, f"Documentation generation failed: {result.stderr}")
+
+        # Optionally, check if the output directory contains the expected files
+        self.assertTrue(os.path.exists(docs_output_dir), "Output directory does not exist")
+        self.assertTrue(os.path.isfile(os.path.join(docs_output_dir, 'index.html')),
+                        "index.html not found in output directory")

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_export.py

@@ -328,9 +328,4 @@ class ModelyExportTest(unittest.TestCase):
         self.test.exportReport()
 
         if os.path.exists(self.test.getWorkspace()):
-            shutil.rmtree(self.test.getWorkspace())
-
-if __name__ == '__main__':
-    unittest.main()
-
-
+            shutil.rmtree(self.test.getWorkspace())

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_export_recurrent.py

@@ -567,7 +567,4 @@ class ModelyExportTest(unittest.TestCase):
         if os.path.exists(test.getWorkspace()):
             shutil.rmtree(test.getWorkspace())
 
-if __name__ == '__main__':
-    unittest.main()
-
 

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_input_dimensions.py

@@ -436,5 +436,3 @@ class ModelyNetworkBuildingTest(unittest.TestCase):
         ## Clip the step to 0 
         test.trainModel(train_dataset='dataset', num_of_epochs=1, train_batch_size=10, step=-4, prediction_samples=20, shuffle_data=True)
         self.assertEqual(8 * 21, len(test.internals.keys()))
-if __name__ == '__main__':
-    unittest.main()

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_json.py

@@ -626,5 +626,3 @@ class ModelyJsonTest(unittest.TestCase):
         self.assertEqual((1, 1, 3), np.array(results['out51']).shape)
 
 
-if __name__ == '__main__':
-    unittest.main()

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_losses.py

@@ -168,7 +168,4 @@ class ModelyTrainingTest(unittest.TestCase):
     #     pass
 
     # def test_analysis_results_connect_state(self):
-    #     pass
-
-if __name__ == '__main__':
-    unittest.main()
+    #     pass

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_model_predict.py

@@ -299,7 +299,7 @@ class ModelyPredictTest(unittest.TestCase):
         #[[0, 5], [1, 2], [2, 7], [3, 3]] * [[-1, -3], [0, 0], [1, -3], [2, 2]]
         self.assertEqual((2,4,2), np.array(results['mul2']).shape)
         self.TestAlmostEqual([[[1, 2], [0, 0], [1, 6], [4, 0]],[[0, -15], [0, 0], [2, -21], [6, 6]]], results['mul2'])
-    
+
     def test_single_in_window_offset_fir(self):
         NeuObj.clearNames()
         # The input must be scalar and the time dimension is compress to 1,
@@ -865,36 +865,37 @@ class ModelyPredictTest(unittest.TestCase):
     def test_initialization(self):
         NeuObj.clearNames()
         input = Input('in1')
-        W = Parameter('W', dimensions=(1,1), init=init_constant)
-        b = Parameter('b', dimensions=1, init=init_constant)
+        W = Parameter('W', dimensions=(1,1), init='init_constant')
+        b = Parameter('b', dimensions=1, init='init_constant')
         o = Output('out', Linear(W=W,b=b)(input.last()))
 
-        W5 = Parameter('W5', dimensions=(1,1), init=init_constant, init_params={'value':5})
-        b2 = Parameter('b2', dimensions=1, init=init_constant, init_params={'value':2})
+        W5 = Parameter('W5', dimensions=(1,1), init='init_constant', init_params={'value':5})
+        b2 = Parameter('b2', dimensions=1, init='init_constant', init_params={'value':2})
         o52 = Output('out52', Linear(W=W5,b=b2)(input.last()))
 
-        par = Parameter('par', dimensions=3, sw=2, init=init_constant)
+        par = Parameter('par', dimensions=3, sw=2, init='init_constant')
         opar = Output('outpar', Fir(W=par)(input.sw(2)))
 
-        par2 = Parameter('par2', dimensions=3, sw=2, init=init_constant, init_params={'value':2})
+        par2 = Parameter('par2', dimensions=3, sw=2, init='init_constant', init_params={'value':2})
         opar2 = Output('outpar2', Fir(W=par2)(input.sw(2)))
 
-        ol = Output('outl', Linear(output_dimension=1,b=True,W_init=init_constant,b_init=init_constant)(input.last()))
-        ol52 = Output('outl52', Linear(output_dimension=1,b=True,W_init=init_constant,b_init=init_constant,W_init_params={'value':5},b_init_params={'value':2})(input.last()))
-        ofpar = Output('outfpar', Fir(output_dimension=3,W_init=init_constant)(input.sw(2)))
-        ofpar2 = Output('outfpar2', Fir(output_dimension=3,W_init=init_constant,W_init_params={'value':2})(input.sw(2)))
+        ol = Output('outl', Linear(output_dimension=1,b=True,W_init='init_constant',b_init='init_constant')(input.last()))
+        ol52 = Output('outl52', Linear(output_dimension=1,b=True,W_init='init_constant',b_init='init_constant',W_init_params={'value':5},b_init_params={'value':2})(input.last()))
+        ofpar = Output('outfpar', Fir(output_dimension=3,W_init='init_constant')(input.sw(2)))
+        ofpar2 = Output('outfpar2', Fir(output_dimension=3,W_init='init_constant',W_init_params={'value':2})(input.sw(2)))
 
-        outnegexp = Output('outnegexp', Fir(output_dimension=3,W_init=init_negexp)(input.sw(2)))
-        outnegexp2 = Output('outnegexp2', Fir(output_dimension=3,W_init=init_negexp,W_init_params={'size_index':1, 'first_value':3, 'lambda':1})(input.sw(2)))
+        outnegexp = Output('outnegexp', Fir(output_dimension=3,W_init='init_negexp')(input.sw(2)))
+        outnegexp2 = Output('outnegexp2', Fir(output_dimension=3,W_init='init_negexp',W_init_params={'size_index':1, 'first_value':3, 'lambda':1})(input.sw(2)))
 
-        outexp = Output('outexp', Fir(output_dimension=3,W_init=init_exp)(input.sw(2)))
-        outexp2 = Output('outexp2', Fir(output_dimension=3,W_init=init_exp,W_init_params={'size_index':1, 'max_value':2, 'lambda':2, 'monotonicity':'increasing'})(input.sw(2)))
+        outexp = Output('outexp', Fir(output_dimension=3,W_init='init_exp')(input.sw(2)))
+        outexp2 = Output('outexp2', Fir(output_dimension=3,W_init='init_exp',W_init_params={'size_index':1, 'max_value':2, 'lambda':2, 'monotonicity':'increasing'})(input.sw(2)))
+        outexp2D = Output('outexp2D', Fir(output_dimension=3,W_init='init_exp',W_init_params={'size_index':1, 'max_value':2, 'lambda':2, 'monotonicity':'decreasing'})(input.sw(2)))
 
-        outlin = Output('outlin', Fir(output_dimension=3,W_init=init_lin)(input.sw(2)))
-        outlin2 = Output('outlin2', Fir(output_dimension=3,W_init=init_lin,W_init_params={'size_index':1, 'first_value':4, 'last_value':5})(input.sw(2)))
+        outlin = Output('outlin', Fir(output_dimension=3,W_init='init_lin')(input.sw(2)))
+        outlin2 = Output('outlin2', Fir(output_dimension=3,W_init='init_lin',W_init_params={'size_index':1, 'first_value':4, 'last_value':5})(input.sw(2)))
 
         n = Modely(visualizer=None,seed=1)
-        n.addModel('model',[o,o52,opar,opar2,ol,ol52,ofpar,ofpar2,outnegexp,outnegexp2,outexp,outexp2,outlin,outlin2])
+        n.addModel('model',[o,o52,opar,opar2,ol,ol52,ofpar,ofpar2,outnegexp,outnegexp2,outexp,outexp2,outexp2D,outlin,outlin2])
         n.neuralizeModel()
         results = n({'in1': [1, 1, 2]})
         self.assertEqual((2,), np.array(results['out']).shape)
@@ -926,6 +927,8 @@ class ModelyPredictTest(unittest.TestCase):
         self.TestAlmostEqual([[[1.0497870445251465,1.0497870445251465,1.0497870445251465]],[[1.099574089050293,1.099574089050293,1.099574089050293]]], results['outexp'])
         self.assertEqual((2,1,3), np.array(results['outexp2']).shape)
         self.TestAlmostEqual([[[0.5413411259651184,1.47151780128479,4]],[[0.81201171875,2.2072768211364746,6]]], results['outexp2'])
+        self.assertEqual((2,1,3), np.array(results['outexp2D']).shape)
+        self.TestAlmostEqual([[[4.0,1.47151780128479,0.5413411259651184]],[[6,2.2072768211364746,0.81201171875]]], results['outexp2D'])
 
         self.assertEqual((2,1,3), np.array(results['outlin']).shape)
         self.TestAlmostEqual([[[1,1,1]],[[1,1,1]]], results['outlin'])
@@ -1370,6 +1373,20 @@ class ModelyPredictTest(unittest.TestCase):
         self.assertEqual((3, 1, 6), np.array(results['out']).shape)
         self.assertEqual([[[1.0, 0.0, 0.0, 0.0, 0.0, 0.0]],[[0.0, 1.0, 0.0, 0.0, 0.0, 0.0]],[[0.0, 0.0, 1.0, 0.0, 0.0, 0.0]]], results['out'])
 
+        def fun(x):
+            import torch
+            return torch.sign(x)
+
+        fuz = Fuzzify(output_dimension=11, range=[-5, 5], functions=[fun,fun])(input.last())
+        out = Output('out2', fuz)
+        test.addModel('out2',[out])
+        test.neuralizeModel()
+        results = test({'in1': [0, 1, 2]})
+        self.assertEqual((3, 1, 11), np.array(results['out2']).shape)
+        self.assertEqual([[[1.0, 1.0, 1.0, 1.0, 1.0, 0.0, -1.0, -1.0, -1.0, -1.0, -1.0]],
+                                [[1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.0, -1.0, -1.0, -1.0, -1.0]],
+                                [[1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.0, -1.0, -1.0, -1.0]]], results['out2'])
+
     def test_sw_on_stream_sw_by_heand(self):
         NeuObj.clearNames()
         input = Input('in1')
@@ -1818,7 +1835,3 @@ class ModelyPredictTest(unittest.TestCase):
         self.TestAlmostEqual(results['in1_sm2'], inin_sm2)
         self.TestAlmostEqual(results['in1_sm2_2'], inin_sm2)
         self.TestAlmostEqual(results['in1_sm2_3'], inin_sm2)
-
-if __name__ == '__main__':
-    unittest.main()
-

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_model_predict_recurrent.py

@@ -1239,7 +1239,3 @@ class ModelyRecurrentPredictTest(unittest.TestCase):
         self.assertEqual((5,), np.array(result['out']).shape)
         self.assertEqual([14.0,3*14+2*3+2*1,50*3+14*2+3*1,181*3+50*2+14*1,0*3+0*2+5*1], result['out'])
 
-
-if __name__ == '__main__':
-    unittest.main()
-

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_network_element.py

@@ -443,7 +443,3 @@ class ModelyNetworkBuildingTest(unittest.TestCase):
         # The output is 2 samples
         self.assertEqual({'out': [1.7170718908309937, 1.9410502910614014]}, example({'x': [-1, 0, 1, 2, 0]}))
         self.assertEqual({'out': [1.7170718908309937, 1.9410502910614014]}, example({'x': [[-1, 0, 1, 2], [0, 1, 2, 0]]}, sampled=True))
-
-
-if __name__ == '__main__':
-    unittest.main()

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_parameters_of_train.py

@@ -25,15 +25,6 @@ def linear_fun(x,a,b):
     return x*a+b
 
 class ModelyTrainingTestParameter(unittest.TestCase):
-    def TestAlmostEqual(self, data1, data2, precision=4):
-        assert np.asarray(data1, dtype=np.float32).ndim == np.asarray(data2, dtype=np.float32).ndim, f'Inputs must have the same dimension! Received {type(data1)} and {type(data2)}'
-        if type(data1) == type(data2) == list:
-            self.assertEqual(len(data1), len(data2))
-            for pred, label in zip(data1, data2):
-                self.TestAlmostEqual(pred, label, precision=precision)
-        else:
-            self.assertAlmostEqual(data1, data2, places=precision)
-
     def test_network_mass_spring_damper(self):
         NeuObj.clearNames()
         x = Input('x')  # Position

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_results.py

@@ -18,15 +18,6 @@ sys.path.append(os.getcwd())
 data_folder = os.path.join(os.path.dirname(__file__), 'data/')
 
 class ModelyTrainingTest(unittest.TestCase):
-    def TestAlmostEqual(self, data1, data2, precision=4):
-        assert np.asarray(data1, dtype=np.float32).ndim == np.asarray(data2, dtype=np.float32).ndim, f'Inputs must have the same dimension! Received {type(data1)} and {type(data2)}'
-        if type(data1) == type(data2) == list:
-            self.assertEqual(len(data1), len(data2))
-            for pred, label in zip(data1, data2):
-                self.TestAlmostEqual(pred, label, precision=precision)
-        else:
-            self.assertAlmostEqual(data1, data2, places=precision)
-
     def test_analysis_results(self):
         NeuObj.clearNames()
         input1 = Input('in1')
@@ -308,12 +299,3 @@ class ModelyTrainingTest(unittest.TestCase):
         self.assertAlmostEqual((np.sum((np.array(A).flatten() - np.array(B).flatten()) ** 2) / 30.0 + np.sum(
             (np.array(C).flatten() - np.array(B).flatten()) ** 2) / 30.0) / 2.0,
                                test.performance['dataset']['total']['mean_error'], places=3)
-
-    # def test_analysis_results_connect(self):
-    #     pass
-
-    # def test_analysis_results_connect_state(self):
-    #     pass
-
-if __name__ == '__main__':
-    unittest.main()

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_train.py

@@ -16,15 +16,6 @@ sys.path.append(os.getcwd())
 data_folder = os.path.join(os.path.dirname(__file__), 'data/')
 
 class ModelyTrainingTest(unittest.TestCase):
-    def TestAlmostEqual(self, data1, data2, precision=4):
-        assert np.asarray(data1, dtype=np.float32).ndim == np.asarray(data2, dtype=np.float32).ndim, f'Inputs must have the same dimension! Received {type(data1)} and {type(data2)}'
-        if type(data1) == type(data2) == list:
-            self.assertEqual(len(data1), len(data2))
-            for pred, label in zip(data1, data2):
-                self.TestAlmostEqual(pred, label, precision=precision)
-        else:
-            self.assertAlmostEqual(data1, data2, places=precision)
-
     def test_training_values_fir(self):
         NeuObj.clearNames()
         input1 = Input('in1')
@@ -280,6 +271,3 @@ class ModelyTrainingTest(unittest.TestCase):
         ## Print the initial weights
         optimizer_defaults = {'weight_decay': 0.3,}
         example.trainModel(train_dataset='dataset', lr=0.01, num_of_epochs=2, optimizer_defaults=optimizer_defaults)
-
-if __name__ == '__main__':
-    unittest.main()

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_train_recurrent.py

@@ -1517,7 +1517,4 @@ class ModelyTrainingTest(unittest.TestCase):
         self.assertListEqual(test2.training['error1']['train'] , test.training['error1']['train'])
         self.assertListEqual(test2.training['error1']['val'], test.training['error1']['val'])
         self.assertListEqual(test2.training['error2']['train'] , test.training['error2']['train'])
-        self.assertListEqual(test2.training['error2']['val'], test.training['error2']['val'])
-
-if __name__ == '__main__':
-    unittest.main()
+        self.assertListEqual(test2.training['error2']['val'], test.training['error2']['val'])

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_utils.py

@@ -23,6 +23,3 @@ class ModelyTrainingTest(unittest.TestCase):
 
         y = linear_interp(x,x_data,y_data)
         self.assertEqual(y.shape, x.shape)  # check that the output has the same shape as the input
-
-if __name__ == '__main__':
-    unittest.main()

{nnodely-1.3.0 → nnodely-1.3.1}/tests/test_visualizer.py

@@ -161,7 +161,4 @@ class ModelyTestVisualizer(unittest.TestCase):
     #     test.loadData(name='dataset', source=dataset)  # Create the dataset
     #     test.trainModel(optimizer='SGD', training_params=params)  # Train the traced model
     #     m.showFunctions(list(test.model_def['Functions'].keys()), xlim=[[-5, 5], [-1, 1]])
-    #     m.closeFunctions()
-
-if __name__ == '__main__':
-    unittest.main()
+    #     m.closeFunctions()