Implement binary search for power start / end search

[Jonas Höchst]

The signal extraction on the spectrogram relies on a linear search.

pyradiotracking/radiotracking/analyze.py

Lines 175 to 196 in a6e953d

 # loop down until threshold is undershot 
 start = ti 
 start_min = 0 if self._spectrogram_last is None else -len(self._spectrogram_last[0]) 
 while start > start_min: 
     if start < 0: 
         power = self._spectrogram_last[fi, start] 
     else: 
         power = fft[start] 
  
     if power < self.signal_threshold: 
         break 
  
     start -= 1 
  
 # loop up until threshold is undershot 
 end = ti 
 while end < len(fft): 
     if fft[end] < self.signal_threshold: 
         ti_skip = end 
         break 
  
     end += 1 

When using binary search, a significant performance gain in low signal thresholds can be expected.