Jump detector v4

analytics/detectors/jump_detector.py

@@ -6,13 +6,15 @@ from scipy.fftpack import fft
 from scipy.signal import argrelextrema
 import math
 
+WINDOW_SIZE = 120
 
 class JumpDetector:
 
     def __init__(self):
         self.segments = []
         self.confidence = 1.5
-        self.convolve_max = 120
+        self.convolve_max = WINDOW_SIZE
+        self.size = 50
     
     async def fit(self, dataframe, segments):
         #self.alpha_finder()
@@ -27,8 +29,7 @@ class JumpDetector:
                 confidences.append(0.20 * (segment_max - segment_min))
                 flat_segment = segment_data.rolling(window=4).mean() #сглаживаем сегмент
                 kde_segment = flat_data.dropna().plot.kde() # distribution density
-                ax = flat_data.dropna().plot.kde()
+                ax_list = kde_segment.get_lines()[0].get_xydata() #take coordinates of kde
-                ax_list = ax.get_lines()[0].get_xydata()
                 mids = argrelextrema(np.array(ax_list), np.less)[0] 
                 maxs = argrelextrema(np.array(ax_list), np.greater)[0]
                 min_peak = maxs[0]
@@ -36,27 +37,15 @@ class JumpDetector:
                 min_line = ax_list[min_peak, 0]
                 max_line = ax_list[max_peak, 0]
                 sigm_heidht = max_line - min_line
-                pat_sigm = utils.logistic_sigmoid(-120, 120, 1, sigm_heidht)
+                pat_sigm = utils.logistic_sigmoid(-WINDOW_SIZE, WINDOW_SIZE, 1, sigm_heidht)
                 for i in range(0, len(pat_sigm)):
                     pat_sigm[i] = pat_sigm[i] + min_line 
-                cen_ind = utils.intersection_segment(flat_segment, mids[0])
+                cen_ind = utils.intersection_segment(flat_segment, mids[0]) #finds all interseprions with median
-                c = []
+                c = [] # choose the correct one interseption by convolve
-                for i in range(len(cen_ind)):
+                jump_center = utils.find_jump_center(cen_ind)
-                    x = cen_ind[i]
+
-                    cx = scipy.signal.fftconvolve(pat_sigm, flat_data[x-120:x+120])
+                segment_cent_index = jump_center - 4
-                    c.append(cx[240])
+                labeled_drop = data[segment_cent_index - WINDOW_SIZE : segment_cent_index + WINDOW_SIZE]
-                
-                # в идеале нужно посмотреть гистограмму сегмента и выбрать среднее значение,
-                # далее от него брать + -120 
-                segment_summ = 0
-                for val in flat_segment:
-                    segment_summ += val
-                segment_mid = segment_summ /  len(flat_segment) #посчитать нормально среднее значение/медиану
-                for ind in range(1, len(flat_segment) - 1):
-                    if flat_segment[ind + 1] > segment_mid and flat_segment[ind - 1] < segment_mid:
-                        flat_mid_index = ind   # найти пересечение средней и графика, получить его индекс
-                segment_mid_index = flat_mid_index - 5
-                labeled_drop = data[segment_mid_index - 120 : segment_mid_index + 120]
                 labeled_min = min(labeled_drop) 
                 for value in labeled_drop: # обрезаем
                     value = value - labeled_min
@@ -65,7 +54,9 @@ class JumpDetector:
                     value = value / labeled_max
                 convolve = scipy.signal.fftconvolve(labeled_drop, labeled_drop)
                 convolve_list.append(max(convolve)) # сворачиваем паттерн
-                # плюс надо впихнуть сюда логистическую сигмоиду и поиск альфы
+                # TODO: add convolve with alpha sigmoid
+                # TODO: add size of jump rize
+
 
         if len(confidences) > 0:
             self.confidence = min(confidences)
@@ -75,7 +66,7 @@ class JumpDetector:
         if len(convolve_list) > 0:
             self.convolve_max = max(convolve_list)
         else:
-            self.convolve_max = 120 # макс метрика свертки равна отступу(120), вау!
+            self.convolve_max = WINDOW_SIZE # макс метрика свертки равна отступу(WINDOW_SIZE), вау!
     
     async def predict(self, dataframe):
         data = dataframe['value']
@@ -91,9 +82,9 @@ class JumpDetector:
         window_size = 24
         all_max_flatten_data = data.rolling(window=window_size).mean()
         all_mins = argrelextrema(np.array(all_max_flatten_data), np.less)[0]        
-        extrema_list = []
+        possible_jumps = utils.find_all_jumps(all_max_flatten_data, 50, self.confidence)
-        # добавить все пересечения экспоненты со сглаженным графиком
 
+        '''
         for i in utils.exponential_smoothing(data + self.confidence, 0.02):
             extrema_list.append(i)
 
@@ -101,8 +92,9 @@ class JumpDetector:
         for i in all_mins:
             if all_max_flatten_data[i] > extrema_list[i]:
                 segments.append(i - window_size)
+        '''
 
-        return [(x - 1, x + 1) for x in self.__filter_prediction(segments, all_max_flatten_data)]
+        return [(x - 1, x + 1) for x in self.__filter_prediction(possible_jumps, all_max_flatten_data)]
 
     def __filter_prediction(self, segments, all_max_flatten_data):
         delete_list = []
@@ -116,10 +108,14 @@ class JumpDetector:
             segments.remove(item)
 
         # изменить секонд делит лист, сделать для свертки с сигмоидой
+        # !!!!!!!!
+        # написать фильтрацию паттернов-джампов! посмотерть каждый сегмент, обрезать его
+        # отнормировать, сравнить с выбранным патерном.
+        # !!!!!!!!
         delete_list = []
-        pattern_data = all_max_flatten_data[segments[0] - 120 : segments[0] + 120]
+        pattern_data = all_max_flatten_data[segments[0] - WINDOW_SIZE : segments[0] + WINDOW_SIZE]
         for segment in segments:
-            convol_data = all_max_flatten_data[segment - 120 : segment + 120]
+            convol_data = all_max_flatten_data[segment - WINDOW_SIZE : segment + WINDOW_SIZE]
             conv = scipy.signal.fftconvolve(pattern_data, convol_data)
             if max(conv) > self.convolve_max * 1.1 or max(conv) < self.convolve_max * 0.9:
                 delete_list.append(segment)

analytics/utils/__init__.py

@@ -61,27 +61,27 @@ def segments_box(segments):
     return min_time, max_time
 
 def intersection_segment(data, median):
+    """
+    Finds all intersections between flatten data and median
+    """
     cen_ind = []
     for i in range(1, len(data)-1):
         if data[i - 1] < median and data[i + 1] > median:
             cen_ind.append(i)
     del_ind = []
-    for i in range(1,len(cen_ind)):
+    for i in range(1, len(cen_ind)):
         if cen_ind[i] == cen_ind[i - 1] + 1:
             del_ind.append(i - 1)
-    del_ind = del_ind[::-1]
-    for i in del_ind:
-        del cen_ind[i]
-    return cen_ind
  
-def logistic_sigmoid(self, x1, x2, alpha, height):
+    return [x for (idx, x) in enumerate(cen_ind) if idx not in del_ind]
-    distribution = []
+
-    for i in range(x1, x2):
+def logistic_sigmoid_distribution(self, x1, x2, alpha, height):
-        F = 1 * height / (1 + math.exp(-i * alpha))
+    return map(lambda x: logistic_sigmoid(x, alpha, height), range(x1, x2))
-        distribution.append(F)
-    return distribution
 
-def findOneJump(data, x, size, height, err):
+def logistic_sigmoid(x, alpha, height):
+    return height / (1 + math.exp(-x * alpha))
+
+def find_one_jump(data, x, size, height, err):
     l = []
     for i in range(x + 1, x + size):
         if (data[i] > data[x] and data[x + size] > data[x] + height):
@@ -91,10 +91,20 @@ def findOneJump(data, x, size, height, err):
     else:
         return 0
 
-def findAllJumps(data, size, height):
+def find_all_jumps(data, size, height):
     possible_jump_list = []
     for i in range(len(data - size)):
-        x = findOneJump(data, i, size, height, 0.9)
+        x = find_one_jump(data, i, size, height, 0.9)
         if x > 0:
             possible_jump_list.append(x)
     return possible_jump_list
+
+def find_jump_center(cen_ind):
+    jump_center = cen_ind[0]
+    for i in range(len(cen_ind)):
+        x = cen_ind[i]
+        cx = scipy.signal.fftconvolve(pat_sigm, flat_data[x - WINDOW_SIZE : x + WINDOW_SIZE])
+        c.append(cx[2 * WINDOW_SIZE])
+        if i > 0 and cx > c[i - 1]:
+            jump_center = x
+    return jump_center