Tried some testing with the standardized version of an algorithm

.gitignore

@@ -10,3 +10,4 @@ __pycache__/
 .cache
 nosetests.xml
 coverage.xml
+*.pyc

src/standardized/ETP_SRI_LinearFitting.py

@@ -20,7 +20,7 @@ class ETP_SRI_LinearFitting(OsipiBase):
 
     # Algorithm requirements
     required_bvalues = 3
-    required_thresholds = [1,1] # Interval from 1 to 1, in case submissions allow a custom number of thresholds
+    required_thresholds = [0,1] # Interval from 1 to 1, in case submissions allow a custom number of thresholds
     required_bounds = False
     required_bounds_optional = True # Bounds may not be required but are optional
     required_initial_guess = False
@@ -45,7 +45,7 @@ def __init__(self, bvalues=None, thresholds=None, bounds=None, initial_guess=Non
         # Check the inputs
 
 
-    def ivim_fit(self, signals, bvalues):
+    def ivim_fit(self, signals, bvalues=None):
         """
             We may want this function to be as simple as
             data and b-values in -> parameter estimates out.
@@ -55,8 +55,13 @@ def ivim_fit(self, signals, bvalues):
             This makes the execution of submissions easy and predictable.
             All execution stepts requires should be performed here.
         """
+        if bvalues is None:
+            bvalues = self.bvalues
 
-        ETP_object = LinearFit(self.thresholds[0])
+        if self.thresholds is None:
+            ETP_object = LinearFit()
+        else:
+            ETP_object = LinearFit(self.thresholds[0])
         f, D, Dstar = ETP_object.ivim_fit(bvalues, signals)
 
 
@@ -73,7 +78,6 @@ def ivim_model(b, S0=1, f=0.1, Dstar=0.03, D=0.001):
 
 signals = ivim_model(bvalues)
 
-model = ETP_SRI_LinearFitting(thresholds=[200])
+model = ETP_SRI_LinearFitting()
 results = model.ivim_fit(signals, bvalues)
-test = model.osipi_simple_bias_and_RMSE_test(SNR=20, bvalues=bvalues, f=0.1, Dstar=0.03, D=0.001, noise_realizations=10)
-#model.print_requirements_osipi()
+test = model.osipi_simple_bias_and_RMSE_test(SNR=20, bvalues=bvalues, f=0.1, Dstar=0.03, D=0.001, noise_realizations=10)

src/wrappers/OsipiBase.py

@@ -6,15 +6,23 @@ class OsipiBase:
     """The base class for OSIPI IVIM fitting"""
 
     def __init__(self, bvalues=None, thresholds=None, bounds=None, initial_guess=None):
-        self.bvalues = np.asarray(bvalues)
-        self.thresholds = np.asarray(thresholds)
-        self.bounds = np.asarray(bounds)
-        self.initial_guess = np.asarray(initial_guess)
+        # Define the attributes as numpy arrays only if they are not None
+        self.bvalues = np.asarray(bvalues) if bvalues is not None else None
+        self.thresholds = np.asarray(thresholds) if thresholds is not None else None
+        self.bounds = np.asarray(bounds) if bounds is not None else None
+        self.initial_guess = np.asarray(initial_guess) if initial_guess is not None else None
 
     def osipi_fit(self, data=None, bvalues=None, initial_guess=None, bounds=None, thresholds=None, **kwargs):
         """Fits the data with the bvalues
         Returns [S0, f, D*, D]
         """
+        if bvalues is None:
+            bvalues = self.bvalues
+
+        args = [data, bvalues, initial_guess, bounds, thresholds]
+        args = [arg for arg in args if arg is not None]
+        f, Dstar, D = self.ivim_fit(*args)
+
         #self.parameter_estimates = self.ivim_fit(data, bvalues)
         pass
 

tests/IVIMmodels/data/test_GenerateData.py

@@ -1,7 +1,7 @@
 import numpy as np
 import numpy.testing as npt
 import pytest
-import torch
+#import torch
 
 from utilities.data_simulation.GenerateData import GenerateData
 

tests/IVIMmodels/unit_tests/test_ivim_fit_linear.py

@@ -7,6 +7,8 @@
 from utilities.data_simulation.GenerateData import GenerateData
 from src.original.ETP_SRI.LinearFitting import LinearFit
 
+# Import the standardized version of the algorithm
+from src.standardized.ETP_SRI_LinearFitting import ETP_SRI_LinearFitting
 
 #run using python -m pytest from the root folder
 
@@ -39,10 +41,17 @@ def test_linear_fit(D, bvals):
 ]
 @pytest.mark.parametrize("f, D, Dp, bvals", test_ivim_data)
 def test_ivim_fit(f, D, Dp, bvals):
+    # We should make a wrapper that runs this for a range of different settings, such as b thresholds, bounds, etc.
+    # An additional inputs to these functions could perhaps be a "settings" class with attributes that are the settings to the
+    # algorithms. I.e. bvalues, thresholds, bounds, initial guesses.
+    # That way, we can write something that defines a range of settings, and then just run them through here.
     gd = GenerateData()
     gd_signal = gd.ivim_signal(D, Dp, f, 1, bvals)
-    fit = LinearFit()
-    [f_fit, D_fit, Dp_fit] = fit.ivim_fit(bvals, gd_signal)
+    #fit = LinearFit()   # This is the old code by ETP
+
+    fit = ETP_SRI_LinearFitting() # This is the standardized format by IAR, which every algorithm will be implemented with
+
+    [f_fit, Dp_fit, D_fit] = fit.ivim_fit(gd_signal, bvals) # Note that I have transposed Dp and D. We should decide on a standard order for these. I usually go with f, Dp, and D ordered after size.
     npt.assert_allclose([f, D], [f_fit, D_fit], atol=1e-5)
     if not np.allclose(f, 0):
         npt.assert_allclose(Dp, Dp_fit, rtol=1e-2, atol=1e-3)