Merge pull request #138 from OSIPI/fix-iar_lu_modified_mix

Update IAR_LU_modified_mix.py

Author: IvanARashid

src/standardized/IAR_LU_modified_mix.py

@@ -93,8 +93,17 @@ def ivim_fit(self, signals, bvalues, **kwargs):
             gtab = gradient_table(bvalues, bvec, b0_threshold=0)
 
             self.IAR_algorithm = IvimModelVP(gtab, bounds=bounds, rescale_results_to_mm2_s=True)
-            
-        fit_results = self.IAR_algorithm.fit(signals)
+
+        try: 
+            fit_results = self.IAR_algorithm.fit(signals)
+        except np.linalg.LinAlgError as err:
+            if 'Singular mtrix' in str(err):
+                # We might have a stochastic error. Try to re-run the fit once.
+                fit_results = self.IAR_algorithm.fit(signals)
+            else:
+                print(str(np.linalg.LinAlgError))
+
+        
 
         #f = fit_results.model_params[1]
         #Dstar = fit_results.model_params[2]