Integrate NNLS reg CV into code structure #44

src/fit/parameters.py

@@ -4,6 +4,7 @@
 import math
 from scipy import signal
 from scipy.sparse import diags
+from scipy.linalg import norm
 from functools import partial
 from typing import Callable
 import json
@@ -595,13 +596,13 @@ def __init__(
         self.tol = tol
         self.reg_order: reg_order
         self.mu = None
-        self.fit_function = Model.NNLSregCV.fit
+        # self.fit_function = Model.NNLSregCV.fit
 
     # @staticmethod
     # def _get_G(basis_CV, H, In, mu, signal_CV):
     #     """Determining lambda function G with cross-validation method."""
-    #     fit = NNLS_reg_fit(basis_CV, H, mu, signal_CV)
-    #     # fit = model.NNLS.fit(basis_CV, H, mu, signal_CV)
+    #     fit = Model.NNLS.fit(1, signal_CV, np.concatenate((basis_CV, mu * H)))
+    #     # fit = NNLS_reg_fit(basis, H, mu, signal)
     #
     #     # Calculating G with CrossValidation method
     #     G = (
@@ -636,23 +637,23 @@ def __init__(
     #     )
     #
     #     # Identity matrix
-    #     In = np.identity(len(signal))  # = 16 = len(b_values)
+    #     In = np.identity(len(self.b_values))
     #
     #     Lambda_left = 0.00001
     #     Lambda_right = 8
     #     midpoint = (Lambda_right + Lambda_left) / 2
     #
     #     # Function (+ delta) and derivative f at left point
-    #     G_left = get_G(basis, H, In, Lambda_left, signal)
-    #     G_leftDiff = get_G(basis, H, In, Lambda_left + tol, signal)
+    #     G_left = self._get_G(basis, H, In, Lambda_left, signal)
+    #     G_leftDiff = self._get_G(basis, H, In, Lambda_left + tol, signal)
     #     f_left = (G_leftDiff - G_left) / tol
     #
     #     count = 0
     #     while abs(Lambda_right - Lambda_left) > tol:
     #         midpoint = (Lambda_right + Lambda_left) / 2
     #         # Function (+ delta) and derivative f at middle point
-    #         G_middle = get_G(basis, H, In, midpoint, signal)
-    #         G_middleDiff = get_G(basis, H, In, midpoint + tol, signal)
+    #         G_middle = self._get_G(basis, H, In, midpoint, signal)
+    #         G_middleDiff = self._get_G(basis, H, In, midpoint + tol, signal)
     #         f_middle = (G_middleDiff - G_middle) / tol
     #
     #         if count > 100: