PlasmaControl · renierts · Jul 20, 2026 · Jul 17, 2026 · Jul 17, 2026 · Jul 17, 2026
diff --git a/BACKEND_REDESIGN_PLAN.md b/BACKEND_REDESIGN_PLAN.md
diff --git a/docs/requirements.txt b/docs/requirements.txt
@@ -1,5 +1,5 @@
-docutils<0.18
-sphinx-rtd-theme==0.5.1
-sphinx-copybutton==0.3.1
+sphinx>=7
+sphinx-rtd-theme>=2.0
+sphinx-copybutton>=0.5
 pyrcn
 scikit-learn
diff --git a/docs/source/api/api.rst b/docs/source/api/api.rst
@@ -6,7 +6,14 @@ PyRCN API
 
    pyrcn
    pyrcn.base
+   pyrcn.nn
    pyrcn.echo_state_network
    pyrcn.extreme_learning_machine
    pyrcn.linear_model
-   pyrcn.cluster
+   pyrcn.model_selection
+   pyrcn.metrics
+   pyrcn.datasets
+   pyrcn.projection
+   pyrcn.preprocessing
+   pyrcn.postprocessing
+   pyrcn.util
diff --git a/docs/source/api/pyrcn.metrics.rst b/docs/source/api/pyrcn.metrics.rst
@@ -0,0 +1,9 @@
+.. _`pyrcn.metrics`:
+
+pyrcn.metrics
+=============
+
+.. automodule:: pyrcn.metrics
+   :members:
+   :undoc-members:
+   :show-inheritance:
diff --git a/docs/source/api/pyrcn.model_selection.rst b/docs/source/api/pyrcn.model_selection.rst
@@ -0,0 +1,9 @@
+.. _`pyrcn.model_selection`:
+
+pyrcn.model_selection
+=====================
+
+.. automodule:: pyrcn.model_selection
+   :members:
+   :undoc-members:
+   :show-inheritance:
diff --git a/docs/source/api/pyrcn.nn.rst b/docs/source/api/pyrcn.nn.rst
@@ -0,0 +1,64 @@
+.. _`pyrcn.nn`:
+
+pyrcn.nn
+========
+
+.. automodule:: pyrcn.nn
+
+Reservoir layers
+----------------
+
+.. autoclass:: pyrcn.nn.Reservoir
+   :members:
+   :undoc-members:
+   :show-inheritance:
+
+.. autoclass:: pyrcn.nn.EulerReservoir
+   :members:
+   :undoc-members:
+   :show-inheritance:
+
+Reservoir cells
+---------------
+
+.. autoclass:: pyrcn.nn.LeakyESNCell
+   :members:
+   :undoc-members:
+   :show-inheritance:
+
+.. autoclass:: pyrcn.nn.EulerESNCell
+   :members:
+   :undoc-members:
+   :show-inheritance:
+
+Feature map and readout
+------------------------
+
+.. autoclass:: pyrcn.nn.InputFeatureMap
+   :members:
+   :undoc-members:
+   :show-inheritance:
+
+.. autoclass:: pyrcn.nn.IncrementalRidge
+   :members:
+   :undoc-members:
+   :show-inheritance:
+
+.. autoclass:: pyrcn.nn.LinearReadout
+   :members:
+   :undoc-members:
+   :show-inheritance:
+
+Gradient training
+-----------------
+
+.. autofunction:: pyrcn.nn.train_readout
+
+.. autofunction:: pyrcn.nn.torch_generator
+
+Weight initializers
+-------------------
+
+.. automodule:: pyrcn.nn.init
+   :members:
+   :undoc-members:
diff --git a/docs/source/api/pyrcn.postprocessing.rst b/docs/source/api/pyrcn.postprocessing.rst
@@ -0,0 +1,9 @@
+.. _`pyrcn.postprocessing`:
+
+pyrcn.postprocessing
+====================
+
+.. automodule:: pyrcn.postprocessing
+   :members:
+   :undoc-members:
+   :show-inheritance:
diff --git a/docs/source/api/pyrcn.preprocessing.rst b/docs/source/api/pyrcn.preprocessing.rst
@@ -0,0 +1,9 @@
+.. _`pyrcn.preprocessing`:
+
+pyrcn.preprocessing
+===================
+
+.. automodule:: pyrcn.preprocessing
+   :members:
+   :undoc-members:
+   :show-inheritance:
diff --git a/docs/source/api/pyrcn.projection.rst b/docs/source/api/pyrcn.projection.rst
@@ -0,0 +1,9 @@
+.. _`pyrcn.projection`:
+
+pyrcn.projection
+================
+
+.. automodule:: pyrcn.projection
+   :members:
+   :undoc-members:
+   :show-inheritance:
diff --git a/docs/source/api/pyrcn.util.rst b/docs/source/api/pyrcn.util.rst
@@ -0,0 +1,9 @@
+.. _`pyrcn.util`:
+
+pyrcn.util
+==========
+
+.. automodule:: pyrcn.util
+   :members:
+   :undoc-members:
+   :show-inheritance:
diff --git a/docs/source/conf.py b/docs/source/conf.py
@@ -40,6 +40,22 @@
 ]
 master_doc = 'index'
 
+# Render numpydoc "Attributes" sections as info fields rather than separate
+# object descriptions (avoids duplicate-object-description warnings for
+# dataclass attributes and properties).
+napoleon_use_ivar = True
+
+# Do not document scikit-learn's dynamically added metadata-routing methods
+# (set_*_request); their docstrings reference sklearn-only glossary terms and
+# labels that do not resolve in this documentation.
+autodoc_default_options = {
+    'exclude-members': (
+        'set_fit_request,set_predict_request,set_partial_fit_request,'
+        'set_score_request,set_transform_request,'
+        'set_predict_proba_request,set_inverse_transform_request'
+    ),
+}
+
 # Add any paths that contain templates here, relative to this directory.
 templates_path = ['_templates']
 

diff --git a/docs/source/getting_started.rst b/docs/source/getting_started.rst
@@ -62,6 +62,7 @@ Building your first Reservoir Computing Network
         >>> from inspect import getmembers, isclass
         >>> getmembers(blocks, isclass)
         [('BatchIntrinsicPlasticity', <class 'pyrcn.base.blocks._input_to_node.BatchIntrinsicPlasticity'>),
+        ('EulerNodeToNode', <class 'pyrcn.base.blocks._node_to_node.EulerNodeToNode'>),
         ('HebbianNodeToNode', <class 'pyrcn.base.blocks._node_to_node.HebbianNodeToNode'>),
         ('InputToNode', <class 'pyrcn.base.blocks._input_to_node.InputToNode'>),
         ('NodeToNode', <class 'pyrcn.base.blocks._node_to_node.NodeToNode'>),
@@ -94,10 +95,12 @@ Training a RCN
 
     To train the ESN, only three steps are required:
 
-    1. Randomly distribute the time-series to each reservoir neuron (**Input-to-Node**).
-    2. Compute the state of each neuron based on the current input and the previous
-    state.
-    2. Compute a linear regression between the reservoir states and the target output.
+    1. Randomly distribute the time-series to each reservoir neuron
+       (**Input-to-Node**).
+    2. Compute the state of each neuron based on the current input and the
+       previous state.
+    3. Compute a linear regression between the reservoir states and the
+       target output.
 
     These steps are handled via :py:func:`pyrcn.echo_state_network.ESNRegressor.fit`,
     which is the most important function to train the ESN model:
@@ -120,6 +123,6 @@ Testing and predict using the ESN
 
     .. doctest::
 
-        >>> y_pred = esn.predict(X[:4000])
+        >>> y_pred = esn.predict(X[:4000].reshape(-1, 1))
 
     .. image:: _static/img/getting_started_mackey_glass_predicted.svg
diff --git a/docs/source/index.rst b/docs/source/index.rst
@@ -33,8 +33,8 @@ It is actively developed to be extended into several directions:
 
 * Interaction with `sktime <https://sktime.org/>`_
 * Interaction with `hmmlearn <https://hmmlearn.readthedocs.io/en/stable/>`_
-* More towards future work: Related architectures, such as Liquid State Machines (LSMs)
-and Perturbative Neural Networks (PNNs)
+* More towards future work: Related architectures, such as Liquid State
+  Machines (LSMs) and Perturbative Neural Networks (PNNs)
 
 PyRCN has successfully been used for several tasks:
 

diff --git a/docs/source/installation.rst b/docs/source/installation.rst
@@ -1,3 +1,5 @@
+.. _installation guide:
+
 ==================
 Installation guide
 ==================
@@ -13,10 +15,10 @@ command in a PowerShell or CommandLine in Windows, or in a shell in Linux/MacOS:
 
     python --version
 
-As any package, **PyRCN** has several dependencies as listed in the `requirements.txt
- <https://github.com/PlasmaControl/PyRCN/blob/main/requirements.txt>`_. To avoid any
-unexpected interaction with the basic system as installed on your computer, we highly
-recommend  using a virtual environment
+As any package, **PyRCN** has several dependencies, declared in
+`pyproject.toml <https://github.com/PlasmaControl/PyRCN/blob/main/pyproject.toml>`_.
+To avoid any unexpected interaction with the basic system as installed on your
+computer, we highly recommend using a virtual environment
 
 You can find more information about virtual environments, by checking the `Python
 documentation on virtual environments and packages

diff --git a/docs/source/introduction.rst b/docs/source/introduction.rst
@@ -1,3 +1,5 @@
+.. _whats rc:
+
 Definition of Reservoir Computing
 =================================
 

diff --git a/examples/MNIST_regressors.ipynb b/examples/MNIST_regressors.ipynb
@@ -21,7 +21,7 @@
     "from sklearn.linear_model import Ridge\n",
     "from sklearn.preprocessing import MinMaxScaler\n",
     "from sklearn.model_selection import RandomizedSearchCV, GridSearchCV, StratifiedKFold, ParameterGrid, cross_validate\n",
-    "from sklearn.utils.fixes import loguniform\n",
+    "from scipy.stats import loguniform\n",
     "from sklearn.metrics import accuracy_score\n",
     "\n",
     "from pyrcn.model_selection import SequentialSearchCV\n",

diff --git a/examples/Video_Door_state_Classification.ipynb b/examples/Video_Door_state_Classification.ipynb
@@ -38,7 +38,7 @@
     "import pandas as pd\n",
     "from tqdm import tqdm\n",
     "import matplotlib.pyplot as plt\n",
-    "from IPython.display import set_matplotlib_formats\n",
+    "from matplotlib_inline.backend_inline import set_matplotlib_formats\n",
     "set_matplotlib_formats('png', 'pdf')\n",
     "from mpl_toolkits.axes_grid1 import make_axes_locatable\n",
     "from matplotlib import ticker\n",
@@ -55,7 +55,7 @@
     "from sklearn.metrics import make_scorer, ConfusionMatrixDisplay\n",
     "from sklearn.model_selection import RandomizedSearchCV, GridSearchCV, ParameterGrid\n",
     "from sklearn.cluster import MiniBatchKMeans\n",
-    "from sklearn.utils.fixes import loguniform\n",
+    "from scipy.stats import loguniform\n",
     "from scipy.stats import uniform\n",
     "from joblib import dump, load\n",
     "\n",
@@ -259,10 +259,10 @@
     "step3_esn_params = {'bias_scaling': np.linspace(0.0, 1.0, 11)}\n",
     "step4_esn_params = {'alpha': loguniform(1e-5, 1e1)}\n",
     "\n",
-    "kwargs_step1 = {'n_iter': 200, 'random_state': 42, 'verbose': 1, 'n_jobs': -1, 'scoring': make_scorer(mean_squared_error, greater_is_better=False, needs_proba=True)}\n",
-    "kwargs_step2 = {'n_iter': 50, 'random_state': 42, 'verbose': 1, 'n_jobs': -1, 'scoring': make_scorer(mean_squared_error, greater_is_better=False, needs_proba=True)}\n",
-    "kwargs_step3 = {'verbose': 1, 'n_jobs': -1, 'scoring': make_scorer(mean_squared_error, greater_is_better=False, needs_proba=True)}\n",
-    "kwargs_step4 = {'n_iter': 50, 'random_state': 42, 'verbose': 1, 'n_jobs': -1, 'scoring': make_scorer(mean_squared_error, greater_is_better=False, needs_proba=True)}\n",
+    "kwargs_step1 = {'n_iter': 200, 'random_state': 42, 'verbose': 1, 'n_jobs': -1, 'scoring': make_scorer(mean_squared_error, greater_is_better=False, response_method='predict_proba')}\n",
+    "kwargs_step2 = {'n_iter': 50, 'random_state': 42, 'verbose': 1, 'n_jobs': -1, 'scoring': make_scorer(mean_squared_error, greater_is_better=False, response_method='predict_proba')}\n",
+    "kwargs_step3 = {'verbose': 1, 'n_jobs': -1, 'scoring': make_scorer(mean_squared_error, greater_is_better=False, response_method='predict_proba')}\n",
+    "kwargs_step4 = {'n_iter': 50, 'random_state': 42, 'verbose': 1, 'n_jobs': -1, 'scoring': make_scorer(mean_squared_error, greater_is_better=False, response_method='predict_proba')}\n",
     "\n",
     "# The searches are defined similarly to the steps of a sklearn.pipeline.Pipeline:\n",
     "searches = [('step1', RandomizedSearchCV, step1_esn_params, kwargs_step1),\n",
@@ -342,7 +342,7 @@
     "                    'leakage': loguniform(1e-5, 1e0),\n",
     "                    'bias_scaling': uniform(loc=0, scale=2)}\n",
     "\n",
-    "kwargs_step0 = {'n_iter': 1000, 'random_state': 42, 'verbose': 1, 'n_jobs': -1, 'scoring': make_scorer(mean_squared_error, greater_is_better=False, needs_proba=True)}\n",
+    "kwargs_step0 = {'n_iter': 1000, 'random_state': 42, 'verbose': 1, 'n_jobs': -1, 'scoring': make_scorer(mean_squared_error, greater_is_better=False, response_method='predict_proba')}\n",
     "\n",
     "base_esn = ESNClassifier(**initially_fixed_params)\n",
     "\n",

diff --git a/examples/Video_Door_state_Classification.py b/examples/Video_Door_state_Classification.py
@@ -12,7 +12,7 @@
     RandomizedSearchCV, GridSearchCV,
     ParameterGrid)
 from sklearn.cluster import MiniBatchKMeans
-from sklearn.utils.fixes import loguniform
+from scipy.stats import loguniform
 from scipy.stats import uniform
 from joblib import dump, load
 

diff --git a/examples/Video_Door_state_Classification_randomized_search.py b/examples/Video_Door_state_Classification_randomized_search.py
@@ -6,8 +6,7 @@
 
 from sklearn.metrics import make_scorer
 from sklearn.model_selection import RandomizedSearchCV
-from sklearn.utils.fixes import loguniform
-from scipy.stats import uniform
+from scipy.stats import loguniform, uniform
 from joblib import dump, load
 
 
@@ -102,7 +101,7 @@ def read_file(fname, Nfr=-1):
 kwargs_step0 = {'n_iter': 1000, 'random_state': 42, 'verbose': 1, 'n_jobs': 1,
                 'scoring': make_scorer(mean_squared_error,
                                        greater_is_better=False,
-                                       needs_proba=True)}
+                                       response_method='predict_proba')}
 
 base_esn = ESNClassifier(**initially_fixed_params)
 

diff --git a/examples/digits-kmeans.py b/examples/digits-kmeans.py
@@ -14,7 +14,7 @@
 from sklearn.decomposition import PCA
 from sklearn.cluster import KMeans
 
-from src.pyrcn.util import get_mnist, tud_colors
+from pyrcn.util import get_mnist
 
 
 matplotlib.rc('image', cmap='binary')
@@ -92,7 +92,7 @@ def main():
                                 labelbottom=False)
 
         ax_barchart.bar(list(map(int, values)), cos_similarity[:, i],
-                        tick_label=values, color=tud_colors['lightblue'])
+                        tick_label=values, color='tab:blue')
         # ax_barchart.set_xlim([0, 9])
         ax_barchart.grid(which='both', axis='y')
         ax_barchart.set_yticks([-1., 0., 1.], minor=False)

diff --git a/examples/digits.ipynb b/examples/digits.ipynb
@@ -520,56 +520,6 @@
     "    print(f\"{esn_cv}\\t{t_fit}\\t{t_inference}\\t{acc_score}\\t{mem_size}\")"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "Alternatively, we can also use a PyTorch implementation of the ESN model"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 13,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from pyrcn.nn import ESN\n",
-    "import torch"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 11,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "base_esn = ESN(input_size=8, hidden_size=50, num_layers=1, nonlinearity='tanh',\n",
-    "               bias=True, input_scaling=0.016952130531190705,\n",
-    "               spectral_radius=1.0214946051551315, bias_scaling=1.500499867548985,\n",
-    "               bias_shift=0., input_sparsity=0.1, recurrent_sparsity=0.1,\n",
-    "               bidirectional=False)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 16,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "torch.Size([8, 50])\n"
-     ]
-    }
-   ],
-   "source": [
-    "for x, y in zip(X_train, y_train):\n",
-    "    x = torch.Tensor(x).float()\n",
-    "    print(base_esn(x)[0].shape)\n",
-    "    break"
-   ]
-  },
   {
    "cell_type": "code",
    "execution_count": null,