Recent changes to corrections.py is dropping peaks too liberally

[dillancellier]

Not sure why yet, but perhaps something to do with the new ordering of the functions in corrections.py

The first peak in this signal will not be returned in the final fit. This isn't because it's not getting a gaussian, but the peak is being dropped after corrections.py :

Code to investigate:

time_range = (-0.2, 0.7)
nlv = 0.1
fs = 1000

# simulate ERP
erp_latency = [0.2, 0.32]
erp_amplitude = [0.5, 0.8]
erp_width = [0.04, 0.04]
erp_params = np.ravel(np.column_stack([erp_latency, erp_amplitude, erp_width]))
np.random.seed(42)
time, erp = simulate_erp(time_range, erp_params, nlv=nlv, fs=fs)

# Apply ERPparam
tfm = ERPparam(verbose=False, max_n_peaks=4)

tfm.fit(time, erp, time_range=[0, 1.0])

[dillancellier]

I can't tell how specific of an issue this is to this signal, though.