Merge pull request #723 from emthompson-usgs/clipfix

Fix bugs in clipping methods

Author: emthompson-usgs

gmprocess/waveform_processing/clipping/clip_detection.py

@@ -39,7 +39,7 @@ def __init__(self, st, test_all=False):
 
     def _clean_trace(self, tr):
         '''
-        Helper function to clean the trace
+        Pre-processing steps.
 
         Args:
             tr (StationTrace):
@@ -49,13 +49,7 @@ def _clean_trace(self, tr):
             clean_tr (StationTrace):
                 Cleaned trace.
         '''
-        t_1 = tr.stats.starttime
-        t_2 = t_1 + 180
-        clean_tr = tr.copy()
-        clean_tr.trim(t_1, t_2)
-        clean_tr.detrend(type='constant')
-        clean_tr.normalize()
-        return clean_tr
+        return tr
 
     def _detect(self, clip_tr):
         '''

gmprocess/waveform_processing/clipping/histogram.py

@@ -239,6 +239,24 @@ def _get_clip_intervals(self, signal, peaks, thresh):
         clip_intervals = self._merge_intervals(clip_intervals, 1)
         return clip_intervals
 
+    def _clean_trace(self, tr):
+        '''
+        Pre-processing steps.
+
+        Args:
+            tr (StationTrace):
+                A single trace in the record.
+
+        Returns:
+            clean_tr (StationTrace):
+                Cleaned trace.
+        '''
+        t_1 = tr.stats.starttime
+        t_2 = t_1 + 180
+        clean_tr = tr.copy()
+        clean_tr.trim(t_1, t_2)
+        return clean_tr
+
     def _detect(self, tr):
         '''
         Test for clipping using the histogram-based method. This is a slight

gmprocess/waveform_processing/clipping/jerk.py

@@ -27,7 +27,7 @@ class Jerk(ClipDetection):
         See parent class.
     '''
 
-    def __init__(self, st, point_thresh=25, test_all=False):
+    def __init__(self, st, point_thresh=400, test_all=False):
         '''
         Constructs all neccessary attributes for the Max_Amp class.
 
@@ -48,6 +48,24 @@ def __init__(self, st, point_thresh=25, test_all=False):
             self.num_outliers = None
         self._get_results()
 
+    def _clean_trace(self, tr):
+        '''
+        Pre-processing steps.
+
+        Args:
+            tr (StationTrace):
+                A single trace in the record.
+
+        Returns:
+            clean_tr (StationTrace):
+                Cleaned trace.
+        '''
+        t_1 = tr.stats.starttime
+        t_2 = t_1 + 180
+        clean_tr = tr.copy()
+        clean_tr.trim(t_1, t_2)
+        return clean_tr
+
     def _detect(self, tr):
         '''
         Check for jerk outliers. Based on method described by:
@@ -65,11 +83,10 @@ def _detect(self, tr):
             bool:
                 Is the trace clipped?
         '''
-        temp_st = self.st
-        temp_st.differentiate()
-        if tr.stats.channel[1] == 'H':
-            temp_st.differentiate()
-        abs_diff = np.abs(tr.data)
+        temp_tr = tr.copy()
+        temp_tr.differentiate()
+        temp_tr.differentiate()
+        abs_diff = np.abs(temp_tr.data)
         median_x100 = 100 * np.median(abs_diff)
         i_jerk, = np.where(abs_diff >= median_x100)
         num_outliers = len(i_jerk)

gmprocess/waveform_processing/clipping/ping.py

@@ -51,6 +51,24 @@ def __init__(self, st, percent_thresh=0.57, test_all=False):
             self.num_outliers = None
         self._get_results()
 
+    def _clean_trace(self, tr):
+        '''
+        Pre-processing steps.
+
+        Args:
+            tr (StationTrace):
+                A single trace in the record.
+
+        Returns:
+            clean_tr (StationTrace):
+                Cleaned trace.
+        '''
+        t_1 = tr.stats.starttime
+        t_2 = t_1 + 180
+        clean_tr = tr.copy()
+        clean_tr.trim(t_1, t_2)
+        return clean_tr
+
     def _detect(self, tr):
         '''
         If any two points differ by more than a threshold, fail the trace.

gmprocess/waveform_processing/clipping/std_dev.py

@@ -65,6 +65,26 @@ def __init__(self, st, amp_thresh=0.85, n_std=12, std_thresh=0.001,
             self.num_outliers = None
         self._get_results()
 
+    def _clean_trace(self, tr):
+        '''
+        Helper function to clean the trace
+
+        Args:
+            tr (StationTrace):
+                A single trace in the record.
+
+        Returns:
+            clean_tr (StationTrace):
+                Cleaned trace.
+        '''
+        t_1 = tr.stats.starttime
+        t_2 = t_1 + 180
+        clean_tr = tr.copy()
+        clean_tr.trim(t_1, t_2)
+        clean_tr.detrend(type='constant')
+        clean_tr.normalize()
+        return clean_tr
+
     def _detect(self, tr):
         '''
         For all points with amplitude greater than amp_thresh, calculate

tests/gmprocess/waveform_processing/clipping/histogram_test.py

@@ -1,7 +1,6 @@
 #!/usr/bin/env python
 # -*- coding: utf-8 -*-
 
-import os
 import numpy as np
 from gmprocess.io.read import read_data
 from gmprocess.io.test_utils import read_data_dir
@@ -25,7 +24,7 @@ def test_num_clip_intervals():
 
     np.testing.assert_equal(
         num_clip_intervals,
-        np.array([0, 9, 37, 10, 17, 7])
+        np.array([0, 9, 37, 10, 16, 8])
     )
 
 
@@ -47,11 +46,11 @@ def test_all_num_clip_intervals():
         num_clip_intervals,
         np.array([[
             0, 0, 0],
-            [0, 9, 17],
+            [0, 9, 18],
             [37, 31, 8],
             [10, 23, 0],
-            [17, 37, 20],
-            [0, 0, 7]
+            [16, 40, 21],
+            [0, 0, 8]
         ])
     )
 

tests/gmprocess/waveform_processing/clipping/jerk_test.py

@@ -25,7 +25,7 @@ def test_num_outliers():
 
     np.testing.assert_equal(
         num_outliers,
-        np.array([1137, 878, 923, 1111, 1025, 1356])
+        np.array([1145, 1227, 872, 860, 926, 1205])
     )
 
 
@@ -46,12 +46,12 @@ def test_all_num_outliers():
     np.testing.assert_equal(
         num_outliers,
         np.array([
-            [0, 1137, 1060],
-            [0, 878, 1298],
-            [0, 923, 806],
-            [0, 1111, 1277],
-            [0, 1025, 1071],
-            [0, 1356, 1542]
+            [1145, 1137, 1158],
+            [1227, 878, 1290],
+            [872, 923, 1158],
+            [860, 1111, 1381],
+            [926, 1025, 954],
+            [1205, 1356, 1600]
         ])
     )