From 669c7cd6109fbf13442b52fe50a6c3f9f0663e61 Mon Sep 17 00:00:00 2001
From: Gianmarco Gurioli <ggurioli@sissa.it>
Date: Fri, 26 Feb 2021 17:20:30 +0100
Subject: [PATCH] Inclusion of ANN method

---
 ezyrb/__init__.py |  3 +-
 ezyrb/ann.py      | 83 +++++++++++++++++++++++++++++++++++++++++++++++
 tests/test_ann.py | 50 ++++++++++++++++++++++++++++
 3 files changed, 135 insertions(+), 1 deletion(-)
 create mode 100755 ezyrb/ann.py
 create mode 100755 tests/test_ann.py

diff --git a/ezyrb/__init__.py b/ezyrb/__init__.py
index 2efc7611..71646f0e 100644
--- a/ezyrb/__init__.py
+++ b/ezyrb/__init__.py
@@ -1,7 +1,7 @@
 __all__ = [
     'database',
     'reduction', 'pod',
-    'approximation', 'rbf', 'linear', 'gpr'
+    'approximation', 'rbf', 'linear', 'gpr', 'ann'
 ]
 
 from .meta import *
@@ -13,3 +13,4 @@
 from .linear import Linear
 from .gpr import GPR
 from .reducedordermodel import ReducedOrderModel
+from .ann import ANN
diff --git a/ezyrb/ann.py b/ezyrb/ann.py
new file mode 100755
index 00000000..c6b2a746
--- /dev/null
+++ b/ezyrb/ann.py
@@ -0,0 +1,83 @@
+"""
+Module for Artificial Neural Network (ANN) Prediction.
+"""
+import torch
+import torch.nn as nn
+import numpy as np
+from .approximation import Approximation
+
+class ANN(Approximation):
+    """
+    Feed-Forward Artifical Neural Network (ANN).
+    : param int trained_epoch: number of already trained iterations.
+    : param criterion: Loss definition (Mean Squared). 
+    : type criterion: torch.nn.modules.loss.MSELoss.
+    
+    Example:
+    >>> import ezyrb
+    >>> import numpy as np
+    >>> x = np.random.uniform(-1, 1, size =(4, 2))
+    >>> y = np.array([np.sin(x[:, 0]), np.cos(x[:, 1]**3)]).T
+    >>> ann = ezyrb.ANN()
+    >>> ann.fit(x, y)
+    >>> y_pred = ann.predict(x)
+    >>> print(y)
+    >>> print(y_pred)
+    """
+        
+    def __init__(self):
+        self.trained_epoch = 0   
+        self.criterion = torch.nn.MSELoss()
+        
+        
+    def fit(self, points, values):
+        """
+        Build the ANN given 'points' and 'values' and perform training.
+        
+        Given the number of neurons per layer, a feed-forward NN is defined. 
+        By default: 
+           - niter, number of training iterations: 20000;
+           - activation function in each inner layer: Tanh; activation function
+             at the output layer: Identity;
+           - optimizer: Adam's method with default parameters 
+             (see, e.g., https://pytorch.org/docs/stable/optim.html);
+           - loss: Mean Squared Loss.
+        
+        :param numpy.ndarray points: the coordinates of the given (training) points.
+        :param numpy.ndarray values: the (training) values in the points.
+        :return the training loss value at termination (after niter iterations).
+        :rtype: float.
+        """
+        layers = [points.shape[1], 10, 5, values.shape[1]]
+        niter = 20000
+        arguments = []
+        for i in range(len(layers)-2):
+            arguments.append(nn.Linear(layers[i], layers[i+1]))
+            arguments.append(nn.Tanh())
+        arguments.append(nn.Linear(layers[len(layers)-2], layers[len(layers)-1]))
+        arguments.append(nn.Identity())
+        self.model = nn.Sequential(*arguments)
+        self.optimizer = torch.optim.Adam(self.model.parameters(), lr=0.001, betas=(0.9, 0.999), eps=1e-08, weight_decay=0, amsgrad=False)
+        points = torch.from_numpy(points).float()
+        values = torch.from_numpy(values).float()
+        for epoch in range(niter):
+            y_pred = self.model(points)
+            loss = self.criterion(y_pred, values)
+            self.optimizer.zero_grad()
+            loss.backward()
+            self.optimizer.step()
+        self.trained_epoch += niter
+        return loss.item()
+    
+    def predict(self, new_point):
+        """
+        Evaluate the ANN at given 'new_points'.
+        
+        :param array_like new_points: the coordinates of the given points.
+        :return: the predicted values via the ANN.
+        :rtype: numpy.ndarray
+        """
+        new_point = np.array(new_point)
+        new_point = torch.from_numpy(new_point).float()
+        y_new = self.model(new_point)
+        return y_new.detach().numpy()
\ No newline at end of file
diff --git a/tests/test_ann.py b/tests/test_ann.py
new file mode 100755
index 00000000..009e614d
--- /dev/null
+++ b/tests/test_ann.py
@@ -0,0 +1,50 @@
+import numpy as np
+import torch.nn as nn
+
+from unittest import TestCase
+from ezyrb import ANN
+
+np.random.seed(17)
+
+def get_xy():
+    npts = 20
+    dinput = 4
+
+    inp = np.random.uniform(-1, 1, size=(npts, dinput))
+    out = np.array([
+        np.sin(inp[:, 0]) + np.sin(inp[:, 1]**2),
+        np.cos(inp[:, 2]) + np.cos(inp[:, 3]**2)
+    ]).T
+
+    return inp, out
+
+class TestANN(TestCase):
+    def test_constructor_empty(self):
+        ann = ANN()
+
+    def test_fit_mono(self):
+        x, y = get_xy()
+        ann = ANN()
+        ann.fit(x[:, 0].reshape(len(x),1), y[:, 0].reshape(len(y),1))
+        assert isinstance(ann.model, nn.Sequential)
+
+    def test_fit(self):
+        x, y = get_xy()
+        ann = ANN()
+        ann.fit(x, y)
+        assert isinstance(ann.model, nn.Sequential)
+
+    def test_predict_01(self):
+        x, y = get_xy()
+        ann = ANN()
+        ann.fit(x, y)
+        test_y = ann.predict(x)
+        np.testing.assert_array_almost_equal(y, test_y, decimal=3)
+
+    def test_predict_02(self):
+        np.random.seed(1)
+        x, y = get_xy()
+        ann = ANN()
+        ann.fit(x, y)
+        test_y = ann.predict(x)
+        np.testing.assert_array_almost_equal(y, test_y, decimal=3)
\ No newline at end of file