Changeset b2b3009d in sasview for src/sas/sasgui/perspectives/fitting

Timestamp:

Apr 4, 2017 12:00:30 PM (8 years ago)

Author:

GitHub <noreply@…>

Branches:

master, ESS_GUI, ESS_GUI_Docs, ESS_GUI_batch_fitting, ESS_GUI_bumps_abstraction, ESS_GUI_iss1116, ESS_GUI_iss879, ESS_GUI_iss959, ESS_GUI_opencl, ESS_GUI_ordering, ESS_GUI_sync_sascalc, costrafo411, magnetic_scatt, release-4.2.2, ticket-1009, ticket-1094-headless, ticket-1242-2d-resolution, ticket-1243, ticket-1249, ticket885, unittest-saveload

Children:

f68d503

Parents:

571bf4b (diff), ed2276f (diff)
Note: this is a merge changeset, the changes displayed below correspond to the merge itself.
Use the (diff) links above to see all the changes relative to each parent.

git-author:

Andrew Jackson <andrew.jackson@…> (04/04/17 12:00:30)

git-committer:

GitHub <noreply@…> (04/04/17 12:00:30)

Message:

Merge pull request #55 from andyfaff/numpy_import

MAINT: import numpy as np

Location:

src/sas/sasgui/perspectives/fitting

Files:

• basepage.py (modified) (13 diffs)
• fitpage.py (modified) (15 diffs)
• fitting.py (modified) (15 diffs)
• model_thread.py (modified) (8 diffs)
• pagestate.py (modified) (4 diffs)
• utils.py (modified) (2 diffs)

src/sas/sasgui/perspectives/fitting/basepage.py

@@ -5 +5 @@
 import os
 import wx
-import numpy
+import numpy as np
 import time
 import copy
@@ -100 +100 @@
         self.graph_id = None
         # Q range for data set
-        self.qmin_data_set = numpy.inf
+        self.qmin_data_set = np.inf
         self.qmax_data_set = None
         self.npts_data_set = 0
@@ -278 +278 @@
 
         """
-        x = numpy.linspace(start=self.qmin_x, stop=self.qmax_x,
+        x = np.linspace(start=self.qmin_x, stop=self.qmax_x,
                            num=self.npts_x, endpoint=True)
         self.data = Data1D(x=x)
@@ -295 +295 @@
         """
         if self.qmin_x >= 1.e-10:
-            qmin = numpy.log10(self.qmin_x)
+            qmin = np.log10(self.qmin_x)
         else:
             qmin = -10.
 
         if self.qmax_x <= 1.e10:
-            qmax = numpy.log10(self.qmax_x)
+            qmax = np.log10(self.qmax_x)
         else:
             qmax = 10.
 
-        x = numpy.logspace(start=qmin, stop=qmax,
+        x = np.logspace(start=qmin, stop=qmax,
                            num=self.npts_x, endpoint=True, base=10.0)
         self.data = Data1D(x=x)
@@ -341 +341 @@
         qstep = self.npts_x
 
-        x = numpy.linspace(start=xmin, stop=xmax, num=qstep, endpoint=True)
-        y = numpy.linspace(start=ymin, stop=ymax, num=qstep, endpoint=True)
+        x = np.linspace(start=xmin, stop=xmax, num=qstep, endpoint=True)
+        y = np.linspace(start=ymin, stop=ymax, num=qstep, endpoint=True)
         # use data info instead
-        new_x = numpy.tile(x, (len(y), 1))
-        new_y = numpy.tile(y, (len(x), 1))
+        new_x = np.tile(x, (len(y), 1))
+        new_y = np.tile(y, (len(x), 1))
         new_y = new_y.swapaxes(0, 1)
         # all data reuire now in 1d array
         qx_data = new_x.flatten()
         qy_data = new_y.flatten()
-        q_data = numpy.sqrt(qx_data * qx_data + qy_data * qy_data)
+        q_data = np.sqrt(qx_data * qx_data + qy_data * qy_data)
         # set all True (standing for unmasked) as default
-        mask = numpy.ones(len(qx_data), dtype=bool)
+        mask = np.ones(len(qx_data), dtype=bool)
         # store x and y bin centers in q space
         x_bins = x
@@ -358 +358 @@
 
         self.data.source = Source()
-        self.data.data = numpy.ones(len(mask))
-        self.data.err_data = numpy.ones(len(mask))
+        self.data.data = np.ones(len(mask))
+        self.data.err_data = np.ones(len(mask))
         self.data.qx_data = qx_data
         self.data.qy_data = qy_data
@@ -783 +783 @@
                     # Skip non-data lines
                     logging.error(traceback.format_exc())
-            return numpy.array(angles), numpy.array(weights)
+            return np.array(angles), np.array(weights)
         except:
             raise
@@ -2107 +2107 @@
         for data in self.data_list:
             # q value from qx and qy
-            radius = numpy.sqrt(data.qx_data * data.qx_data +
+            radius = np.sqrt(data.qx_data * data.qx_data +
                                 data.qy_data * data.qy_data)
             # get unmasked index
@@ -2113 +2113 @@
                          (radius <= float(self.qmax.GetValue()))
             index_data = (index_data) & (data.mask)
-            index_data = (index_data) & (numpy.isfinite(data.data))
+            index_data = (index_data) & (np.isfinite(data.data))
 
             if len(index_data[index_data]) < 10:
@@ -2148 +2148 @@
             index_data = (float(self.qmin.GetValue()) <= radius) & \
                          (radius <= float(self.qmax.GetValue()))
-            index_data = (index_data) & (numpy.isfinite(data.y))
+            index_data = (index_data) & (np.isfinite(data.y))
 
             if len(index_data[index_data]) < 5:
@@ -2220 +2220 @@
 
                 # Check that min is less than max
-                low = -numpy.inf if min_str == "" else float(min_str)
-                high = numpy.inf if max_str == "" else float(max_str)
+                low = -np.inf if min_str == "" else float(min_str)
+                high = np.inf if max_str == "" else float(max_str)
                 if high < low:
                     min_ctrl.SetBackgroundColour("pink")
@@ -2631 +2631 @@
             self.qmin_x = data_min
             self.qmax_x = math.sqrt(x * x + y * y)
-            # self.data.mask = numpy.ones(len(self.data.data),dtype=bool)
+            # self.data.mask = np.ones(len(self.data.data),dtype=bool)
             # check smearing
             if not self.disable_smearer.GetValue():
@@ -3343 +3343 @@
 
             if value[1] == 'array':
-                pd_vals = numpy.array(value[2])
-                pd_weights = numpy.array(value[3])
+                pd_vals = np.array(value[2])
+                pd_weights = np.array(value[3])
                 if len(pd_vals) == 0 or len(pd_vals) != len(pd_weights):
                     msg = ("bad array distribution parameters for %s"

src/sas/sasgui/perspectives/fitting/fitpage.py

@@ -6 +6 @@
 import wx
 import wx.lib.newevent
-import numpy
+import numpy as np
 import copy
 import math
@@ -1115 +1115 @@
             if item.GetValue():
                 if button_list.index(item) == 0:
-                    flag = 0  # dy = numpy.ones_like(dy_data)
+                    flag = 0  # dy = np.ones_like(dy_data)
                 elif button_list.index(item) == 1:
                     flag = 1  # dy = dy_data
                 elif button_list.index(item) == 2:
-                    flag = 2  # dy = numpy.sqrt(numpy.abs(data))
+                    flag = 2  # dy = np.sqrt(np.abs(data))
                 elif button_list.index(item) == 3:
-                    flag = 3  # dy = numpy.abs(data)
+                    flag = 3  # dy = np.abs(data)
                 break
         return flag
@@ -1422 +1422 @@
         key = event.GetKeyCode()
         length = len(self.data.x)
-        indx = (numpy.abs(self.data.x - x_data)).argmin()
+        indx = (np.abs(self.data.x - x_data)).argmin()
         # return array.flat[idx]
         if key == wx.WXK_PAGEUP or key == wx.WXK_NUMPAD_PAGEUP:
@@ -1477 +1477 @@
                     self.enable2D:
                 # set mask
-                radius = numpy.sqrt(self.data.qx_data * self.data.qx_data +
+                radius = np.sqrt(self.data.qx_data * self.data.qx_data +
                                     self.data.qy_data * self.data.qy_data)
                 index_data = ((self.qmin_x <= radius) & (radius <= self.qmax_x))
                 index_data = (index_data) & (self.data.mask)
-                index_data = (index_data) & (numpy.isfinite(self.data.data))
+                index_data = (index_data) & (np.isfinite(self.data.data))
                 if len(index_data[index_data]) < 10:
                     msg = "Cannot Plot :No or too little npts in"
@@ -1598 +1598 @@
                 and data.dqx_data.any() != 0:
                 self.smear_type = "Pinhole2d"
-                self.dq_l = format_number(numpy.average(data.dqx_data))
-                self.dq_r = format_number(numpy.average(data.dqy_data))
+                self.dq_l = format_number(np.average(data.dqx_data))
+                self.dq_r = format_number(np.average(data.dqy_data))
                 return
             else:
                 return
         # check if it is pinhole smear and get min max if it is.
-        if data.dx is not None and numpy.any(data.dx):
+        if data.dx is not None and np.any(data.dx):
             self.smear_type = "Pinhole"
             self.dq_l = data.dx[0]
@@ -1612 +1612 @@
         elif data.dxl is not None or data.dxw is not None:
             self.smear_type = "Slit"
-            if data.dxl is not None and numpy.all(data.dxl, 0):
+            if data.dxl is not None and np.all(data.dxl, 0):
                 self.dq_l = data.dxl[0]
-            if data.dxw is not None and numpy.all(data.dxw, 0):
+            if data.dxw is not None and np.all(data.dxw, 0):
                 self.dq_r = data.dxw[0]
         # return self.smear_type,self.dq_l,self.dq_r
@@ -1914 +1914 @@
                 self.default_mask = copy.deepcopy(self.data.mask)
                 if self.data.err_data is not None \
-                        and numpy.any(self.data.err_data):
+                        and np.any(self.data.err_data):
                     di_flag = True
                 if self.data.dqx_data is not None \
-                        and numpy.any(self.data.dqx_data):
+                        and np.any(self.data.dqx_data):
                     dq_flag = True
             else:
                 self.slit_smearer.Enable(True)
                 self.pinhole_smearer.Enable(True)
-                if self.data.dy is not None and numpy.any(self.data.dy):
+                if self.data.dy is not None and np.any(self.data.dy):
                     di_flag = True
-                if self.data.dx is not None and numpy.any(self.data.dx):
+                if self.data.dx is not None and np.any(self.data.dx):
                     dq_flag = True
-                elif self.data.dxl is not None and numpy.any(self.data.dxl):
+                elif self.data.dxl is not None and np.any(self.data.dxl):
                     dq_flag = True
 
@@ -2060 +2060 @@
         if self.data.__class__.__name__ == "Data2D" or \
                         self.enable2D:
-            radius = numpy.sqrt(self.data.qx_data * self.data.qx_data +
+            radius = np.sqrt(self.data.qx_data * self.data.qx_data +
                                 self.data.qy_data * self.data.qy_data)
             index_data = (self.qmin_x <= radius) & (radius <= self.qmax_x)
             index_data = (index_data) & (self.data.mask)
-            index_data = (index_data) & (numpy.isfinite(self.data.data))
+            index_data = (index_data) & (np.isfinite(self.data.data))
             npts2fit = len(self.data.data[index_data])
         else:
@@ -2097 +2097 @@
         # make sure stop button to fit button all the time
         self._on_fit_complete()
-        if out is None or not numpy.isfinite(chisqr):
+        if out is None or not np.isfinite(chisqr):
             raise ValueError, "Fit error occured..."
 
@@ -2108 +2108 @@
 
         # Check if chi2 is finite
-        if chisqr is not None and numpy.isfinite(chisqr):
+        if chisqr is not None and np.isfinite(chisqr):
             # format chi2
             chi2 = format_number(chisqr, True)
@@ -2160 +2160 @@
 
                         if cov[ind] is not None:
-                            if numpy.isfinite(float(cov[ind])):
+                            if np.isfinite(float(cov[ind])):
                                 val_err = format_number(cov[ind], True)
                                 item[4].SetForegroundColour(wx.BLACK)
@@ -2280 +2280 @@
             self.smear_type = 'Pinhole2d'
             len_data = len(data.data)
-            data.dqx_data = numpy.zeros(len_data)
-            data.dqy_data = numpy.zeros(len_data)
+            data.dqx_data = np.zeros(len_data)
+            data.dqy_data = np.zeros(len_data)
         else:
             self.smear_type = 'Pinhole'
             len_data = len(data.x)
-            data.dx = numpy.zeros(len_data)
+            data.dx = np.zeros(len_data)
             data.dxl = None
             data.dxw = None
@@ -2458 +2458 @@
         try:
             self.dxl = float(self.smear_slit_height.GetValue())
-            data.dxl = self.dxl * numpy.ones(data_len)
+            data.dxl = self.dxl * np.ones(data_len)
             self.smear_slit_height.SetBackgroundColour(wx.WHITE)
         except:
             self.dxl = None
-            data.dxl = numpy.zeros(data_len)
+            data.dxl = np.zeros(data_len)
             if self.smear_slit_height.GetValue().lstrip().rstrip() != "":
                 self.smear_slit_height.SetBackgroundColour("pink")
@@ -2471 +2471 @@
             self.dxw = float(self.smear_slit_width.GetValue())
             self.smear_slit_width.SetBackgroundColour(wx.WHITE)
-            data.dxw = self.dxw * numpy.ones(data_len)
+            data.dxw = self.dxw * np.ones(data_len)
         except:
             self.dxw = None
-            data.dxw = numpy.zeros(data_len)
+            data.dxw = np.zeros(data_len)
             if self.smear_slit_width.GetValue().lstrip().rstrip() != "":
                 self.smear_slit_width.SetBackgroundColour("pink")
@@ -2601 +2601 @@
             if event is None:
                 output = "-"
-            elif not numpy.isfinite(event.output):
+            elif not np.isfinite(event.output):
                 output = "-"
             else:

src/sas/sasgui/perspectives/fitting/fitting.py

@@ -16 +16 @@
 import wx
 import logging
-import numpy
+import numpy as np
 import time
 from copy import deepcopy
@@ -1323 +1323 @@
                     new_theory = copy_data.data
                     new_theory[res.index] = res.theory
-                    new_theory[res.index == False] = numpy.nan
+                    new_theory[res.index == False] = np.nan
                     correct_result = True
                 #get all fittable parameters of the current model
@@ -1332 +1332 @@
                         param_list.remove(param)
                 if not correct_result or res.fitness is None or \
-                    not numpy.isfinite(res.fitness) or \
-                    numpy.any(res.pvec == None) or not \
-                    numpy.all(numpy.isfinite(res.pvec)):
+                    not np.isfinite(res.fitness) or \
+                        np.any(res.pvec == None) or not \
+                        np.all(np.isfinite(res.pvec)):
                     data_name = str(None)
                     if data is not None:
@@ -1343 +1343 @@
                     msg += "Data %s and Model %s did not fit.\n" % (data_name,
                                                                     model_name)
-                    ERROR = numpy.NAN
+                    ERROR = np.NAN
                     cell = BatchCell()
                     cell.label = res.fitness
@@ -1357 +1357 @@
                             batch_inputs["error on %s" % str(param)].append(ERROR)
                 else:
-                    # TODO: Why sometimes res.pvec comes with numpy.float64?
+                    # TODO: Why sometimes res.pvec comes with np.float64?
                     # probably from scipy lmfit
-                    if res.pvec.__class__ == numpy.float64:
+                    if res.pvec.__class__ == np.float64:
                         res.pvec = [res.pvec]
 
@@ -1523 +1523 @@
                 fit_msg = res.mesg
                 if res.fitness is None or \
-                    not numpy.isfinite(res.fitness) or \
-                    numpy.any(res.pvec == None) or \
-                    not numpy.all(numpy.isfinite(res.pvec)):
+                    not np.isfinite(res.fitness) or \
+                        np.any(res.pvec == None) or \
+                    not np.all(np.isfinite(res.pvec)):
                     fit_msg += "\nFitting did not converge!!!"
                     wx.CallAfter(self._update_fit_button, page_id)
                 else:
                     #set the panel when fit result are float not list
-                    if res.pvec.__class__ == numpy.float64:
+                    if res.pvec.__class__ == np.float64:
                         pvec = [res.pvec]
                     else:
                         pvec = res.pvec
-                    if res.stderr.__class__ == numpy.float64:
+                    if res.stderr.__class__ == np.float64:
                         stderr = [res.stderr]
                     else:
@@ -1682 +1682 @@
         if dy is None:
             new_plot.is_data = False
-            new_plot.dy = numpy.zeros(len(y))
+            new_plot.dy = np.zeros(len(y))
             # If this is a theory curve, pick the proper symbol to make it a curve
             new_plot.symbol = GUIFRAME_ID.CURVE_SYMBOL_NUM
@@ -1731 +1731 @@
         """
         try:
-            numpy.nan_to_num(y)
+            np.nan_to_num(y)
             new_plot = self.create_theory_1D(x, y, page_id, model, data, state,
                                              data_description=model.name,
@@ -1815 +1815 @@
         that can be plot.
         """
-        numpy.nan_to_num(image)
+        np.nan_to_num(image)
         new_plot = Data2D(image=image, err_image=data.err_data)
         new_plot.name = model.name + '2d'
@@ -2007 +2007 @@
         if data_copy.__class__.__name__ == "Data2D":
             if index == None:
-                index = numpy.ones(len(data_copy.data), dtype=bool)
+                index = np.ones(len(data_copy.data), dtype=bool)
             if weight != None:
                 data_copy.err_data = weight
             # get rid of zero error points
             index = index & (data_copy.err_data != 0)
-            index = index & (numpy.isfinite(data_copy.data))
+            index = index & (np.isfinite(data_copy.data))
             fn = data_copy.data[index]
             theory_data = self.page_finder[page_id].get_theory_data(fid=data_copy.id)
@@ -2022 +2022 @@
             # 1 d theory from model_thread is only in the range of index
             if index == None:
-                index = numpy.ones(len(data_copy.y), dtype=bool)
+                index = np.ones(len(data_copy.y), dtype=bool)
             if weight != None:
                 data_copy.dy = weight
             if data_copy.dy == None or data_copy.dy == []:
-                dy = numpy.ones(len(data_copy.y))
+                dy = np.ones(len(data_copy.y))
             else:
                 ## Set consistently w/AbstractFitengine:
@@ -2047 +2047 @@
             return
 
-        residuals = res[numpy.isfinite(res)]
+        residuals = res[np.isfinite(res)]
         # get chisqr only w/finite
-        chisqr = numpy.average(residuals * residuals)
+        chisqr = np.average(residuals * residuals)
 
         self._plot_residuals(page_id=page_id, data=data_copy,
@@ -2086 +2086 @@
             residuals.qy_data = data_copy.qy_data
             residuals.q_data = data_copy.q_data
-            residuals.err_data = numpy.ones(len(residuals.data))
+            residuals.err_data = np.ones(len(residuals.data))
             residuals.xmin = min(residuals.qx_data)
             residuals.xmax = max(residuals.qx_data)
@@ -2100 +2100 @@
             # 1 d theory from model_thread is only in the range of index
             if data_copy.dy == None or data_copy.dy == []:
-                dy = numpy.ones(len(data_copy.y))
+                dy = np.ones(len(data_copy.y))
             else:
                 if weight == None:
-                    dy = numpy.ones(len(data_copy.y))
+                    dy = np.ones(len(data_copy.y))
                 ## Set consitently w/AbstractFitengine:
                 ## But this should be corrected later.
@@ -2122 +2122 @@
                 residuals.y = (fn - gn[index]) / en
             residuals.x = data_copy.x[index]
-            residuals.dy = numpy.ones(len(residuals.y))
+            residuals.dy = np.ones(len(residuals.y))
             residuals.dx = None
             residuals.dxl = None

src/sas/sasgui/perspectives/fitting/model_thread.py

@@ -4 +4 @@
 
 import time
-import numpy
+import numpy as np
 import math
 from sas.sascalc.data_util.calcthread import CalcThread
@@ -68 +68 @@
 
         # Define matrix where data will be plotted
-        radius = numpy.sqrt((self.data.qx_data * self.data.qx_data) + \
+        radius = np.sqrt((self.data.qx_data * self.data.qx_data) + \
                     (self.data.qy_data * self.data.qy_data))
 
@@ -75 +75 @@
         index_model = (self.qmin <= radius) & (radius <= self.qmax)
         index_model = index_model & self.data.mask
-        index_model = index_model & numpy.isfinite(self.data.data)
+        index_model = index_model & np.isfinite(self.data.data)
 
         if self.smearer is not None:
@@ -91 +91 @@
                 self.data.qy_data[index_model]
             ])
-        output = numpy.zeros(len(self.data.qx_data))
+        output = np.zeros(len(self.data.qx_data))
         # output default is None
         # This method is to distinguish between masked
@@ -163 +163 @@
         """
         self.starttime = time.time()
-        output = numpy.zeros((len(self.data.x)))
+        output = np.zeros((len(self.data.x)))
         index = (self.qmin <= self.data.x) & (self.data.x <= self.qmax)
 
@@ -175 +175 @@
                                                              self.qmax)
             mask = self.data.x[first_bin:last_bin+1]
-            unsmeared_output = numpy.zeros((len(self.data.x)))
+            unsmeared_output = np.zeros((len(self.data.x)))
             unsmeared_output[first_bin:last_bin+1] = self.model.evalDistribution(mask)
             self.smearer.model = self.model
@@ -183 +183 @@
             # Check that the arrays are compatible. If we only have a model but no data,
             # the length of data.y will be zero.
-            if isinstance(self.data.y, numpy.ndarray) and output.shape == self.data.y.shape:
-                unsmeared_data = numpy.zeros((len(self.data.x)))
-                unsmeared_error = numpy.zeros((len(self.data.x)))
+            if isinstance(self.data.y, np.ndarray) and output.shape == self.data.y.shape:
+                unsmeared_data = np.zeros((len(self.data.x)))
+                unsmeared_error = np.zeros((len(self.data.x)))
                 unsmeared_data[first_bin:last_bin+1] = self.data.y[first_bin:last_bin+1]\
                                                         * unsmeared_output[first_bin:last_bin+1]\
@@ -209 +209 @@
 
         if p_model is not None and s_model is not None:
-            sq_values = numpy.zeros((len(self.data.x)))
-            pq_values = numpy.zeros((len(self.data.x)))
+            sq_values = np.zeros((len(self.data.x)))
+            pq_values = np.zeros((len(self.data.x)))
             sq_values[index] = s_model.evalDistribution(self.data.x[index])
             pq_values[index] = p_model.evalDistribution(self.data.x[index])

src/sas/sasgui/perspectives/fitting/pagestate.py

@@ -18 +18 @@
 import copy
 import logging
-import numpy
+import numpy as np
 import traceback
 
@@ -410 +410 @@
         for fittable, name, value, _, uncert, lower, upper, units in params:
             if not value:
-                value = numpy.nan
+                value = np.nan
             if not uncert or uncert[1] == '' or uncert[1] == 'None':
                 uncert[0] = False
-                uncert[1] = numpy.nan
+                uncert[1] = np.nan
             if not upper or upper[1] == '' or upper[1] == 'None':
                 upper[0] = False
-                upper[1] = numpy.nan
+                upper[1] = np.nan
             if not lower or lower[1] == '' or lower[1] == 'None':
                 lower[0] = False
-                lower[1] = numpy.nan
+                lower[1] = np.nan
             if is_string:
                 p[name] = str(value)
@@ -449 +449 @@
                 lower = params.get(name + ".lower", '-inf')
                 units = params.get(name + ".units")
-                if std is not None and std is not numpy.nan:
+                if std is not None and std is not np.nan:
                     std = [True, str(std)]
                 else:
                     std = [False, '']
-                if lower is not None and lower is not numpy.nan:
+                if lower is not None and lower is not np.nan:
                     lower = [True, str(lower)]
                 else:
                     lower = [True, '-inf']
-                if upper is not None and upper is not numpy.nan:
+                if upper is not None and upper is not np.nan:
                     upper = [True, str(upper)]
                 else:
@@ -1100 +1100 @@
                                        % (line, tagname, name))
                                 logging.error(msg + traceback.format_exc())
-                        dic[name] = numpy.array(value_list)
+                        dic[name] = np.array(value_list)
                     setattr(self, varname, dic)
 

src/sas/sasgui/perspectives/fitting/utils.py

@@ -2 +2 @@
 Module contains functions frequently used in this package
 """
-import numpy
+import numpy as np
 
 
@@ -19 +19 @@
         data = data.y
     if flag == 0:
-        weight = numpy.ones_like(data)
+        weight = np.ones_like(data)
     elif flag == 1:
         weight = dy_data
     elif flag == 2:
-        weight = numpy.sqrt(numpy.abs(data))
+        weight = np.sqrt(np.abs(data))
     elif flag == 3:
-        weight = numpy.abs(data)
+        weight = np.abs(data)
     return weight