Hello, we are trying to look at the spike detection analysis in one part of our research work. We faced some issues and would like to get your opinion and recommendation on it

"from scipy import signal, signal.triang(np.round(down_samp_freq * 0.06)) (under detect function)" doesn't work causing error - module scipy.signal has no attribute triang.

So in 1.13.1 (not stable) version, using

from scipy.signal.windows import triang
template= triang(np.round(down_samp_freq*0.06))
kernel = np.array([-2, -1, 1, 2]) / float(8)
template = np.convolve(kernel, np.convolve(template, kernel, 'valid') ,'full'),

we got a signal as attached

But as it is mentined in webpage of scipy to shift to stable version, we decided to use sawtooth (which was mentioned that it will return triangular waveform) to recreate the template waveform

t= np.linspace(0,1,12)
template=signal.sawtooth(2np.pit)
kernel = np.array([-2, -1, 1, 2]) / float(8)
template = np.convolve(kernel, np.convolve(template, kernel, 'valid') ,'full')

and we got image like this

Also, in both figures- time and amplitude varies. If you could provide us image or values(both amplitude and time) that u obtained for template matching , we could try to change it accordingly.