Merge pull request #261 from fitbenchmarking/242_sorting

Moving data sorting from plotting to data input

(Merging as admin as Tyrone is away, but his requested changes were fixed)

fitbenchmarking/fitbenchmark_one_problem.py

@@ -131,10 +131,12 @@ def benchmark(controller, minimizers):
 
         if chi_sq < min_chi_sq:
             min_chi_sq = chi_sq
+            index = controller.sorted_index
             best_fit = plot_helper.data(name=minimizer,
                                         x=controller.data_x,
                                         y=controller.results,
-                                        E=controller.data_e)
+                                        E=controller.data_e,
+                                        sorted_index=index)
 
         individual_result = \
             misc.create_result_entry(problem=controller.problem,

fitbenchmarking/fitting/controllers/base_controller.py

@@ -87,6 +87,9 @@ def _correct_data(self):
             if self.data_e is not None:
                 self.data_e = self.data_e[mask]
 
+        # Stores the indices of the sorted data
+        self.sorted_index = np.argsort(self.data_x)
+
     def prepare(self):
         """
         Check that function and minimizer have been set.

fitbenchmarking/fitting/plotting/plot_helper.py

@@ -19,12 +19,13 @@ class data:
     Holds all the data that is used in the plotting process.
     """
 
-    def __init__(self, name=None, x=[], y=[], E=[]):
+    def __init__(self, name=None, x=[], y=[], E=[], sorted_index=None):
         """
         Creates a data object.
         @param x :: the x data
         @param y :: the y data
         @param E :: the (y) errors
+        @param sorted_index :: the sorted indices from the x data
         """
 
         if name is not None:
@@ -39,28 +40,19 @@ def __init__(self, name=None, x=[], y=[], E=[]):
             self.E = np.zeros(len(self.x))
         else:
             self.E = copy.copy(E)
+
+        if sorted_index is not None:
+            self.x = self.x[sorted_index]
+            self.y = self.y[sorted_index]
+            self.z = self.E[sorted_index]
+
         self.showError = False
         self.markers = "x"
         self.colour = "k"
         self.linestyle = '--'
         self.z_order = 1
         self.linewidth = 1
 
-    def order_data(self):
-        """
-        Ensures that the data is in ascending order in x.
-        Prevents line plots looping back on themselves.
-        """
-
-        xData = self.x
-        yData = self.y
-        eData = self.E
-
-        index = np.argsort(xData)
-        xData = xData[index]
-        yData = yData[index]
-        eData = eData[index]
-
 
 class plot(data):
     """
@@ -96,7 +88,6 @@ def make_scatter_plot(self, save=""):
         self.set_plot_misc()
         self.save_plot(save)
 
-
         plt.close()
 
     def set_plot_misc(self):

fitbenchmarking/fitting/plotting/plots.py

@@ -98,7 +98,6 @@ def make_best_fit_plot(name, raw_data, best_fit, count, figures_dir):
     best_fit.colour = 'lime'
     best_fit.zorder = 2
     best_fit.linewidth = 1.5
-    best_fit.order_data()
     fig.add_data(best_fit)
     fig.labels['y'] = "Arbitrary units"
     fig.labels['x'] = "Time ($\mu s$)"
@@ -126,7 +125,6 @@ def make_starting_guess_plot(raw_data, problem, count, figures_dir):
     xData = problem.data_x
     yData = problem.eval_starting_params(0)
     startData = data("Start Guess", xData, yData)
-    startData.order_data()
     startData.colour = "red"
     startData.markers = ''
     startData.linestyle = "-"

fitbenchmarking/parsing/base_fitting_problem.py

@@ -7,6 +7,7 @@
     # python3
     from itertools import zip_longest as izip_longest
 
+
 class BaseFittingProblem:
     """
     Definition of a base class implementation of the fitting test problem,

fitbenchmarking/parsing/parse.py

@@ -21,7 +21,7 @@ def parse_problem_file(prob_file):
 
     prob_type = determine_problem_type(prob_file)
     logger.info("Loading {0} formatted problem definition file {1} | Path: "
-                "{2}".format(prob_type,os.path.basename(prob_file),prob_file[prob_file.find('fitbenchmarking'):]))
+                "{2}".format(prob_type, os.path.basename(prob_file), prob_file[prob_file.find('fitbenchmarking'):]))
 
     if prob_type == "NIST":