feat: add db4 wavelet func

Handle_emg_data.py

@@ -44,7 +44,7 @@ def get_emg_str(emg_nr):
 
 def get_min_max_timestamp(df:DataFrame):
     min = int(np.floor(df['timestamp'].min()))
-    max = int(np.ceil(df['timestamp'].max()))
+    max = df['timestamp'].max()
     return min, max
 
 

Signal_prep.py

@@ -2,6 +2,8 @@ import numpy as np
 import matplotlib.pyplot as plt
 from pandas.core.frame import DataFrame
 from scipy.fft import fft, fftfreq
+import pywt
+from scipy.signal import wavelets
 
 import Handle_emg_data as Handler
 
@@ -97,31 +99,38 @@ def load_user_emg_data():
     return csv_handler.data_container_dict
 
 # Takes in a df and outputs np arrays for x and y values
-def prep_df_for_trans(df:DataFrame):
+def prep_df(df:DataFrame):
-    sample_rate = SAMPLE_RATE
     min, duration = Handler.get_min_max_timestamp(df)
-    x = np.linspace(0, duration, SAMPLE_RATE * duration, endpoint=False)
     y = df.iloc[:,1].to_numpy()
-    return x, y, duration
+    return y, duration
 
 def normalize_wave(y_values):
     y = np.int16((y_values / y_values.max()) * 32767)
     return y
 
 
-def transformed_df(df:DataFrame):
+def fft_of_df(df:DataFrame):
-    x_values, y_values, duration = prep_df_for_trans(df)
+    y_values, duration = prep_df(df)
-    N = SAMPLE_RATE * duration
+    N = y_values.size
     norm = normalize_wave(y_values)
     x_f = fftfreq(N, 1 / SAMPLE_RATE)
     y_f = fft(norm)
-    return x_f, y_f
+    return x_f, y_f, duration
+
+def denoise_signal_pywt(df:DataFrame):
+    y_values, duration = prep_df(df)
+    norm = normalize_wave(y_values)
+    wavelet = pywt.Wavelet('db4')
+    cA, cD = pywt.dwt(norm, wavelet)
+    x =  np.array(range(int(np.floor((y_values.size + wavelet.dec_len - 1) / 2))))
+    print(x)
+    return x, cA 
 
 def plot_df(df:DataFrame):
     lines = df.plot.line(x='timestamp')
     plt.show()
 
-def plot_fft(x_f, y_f):
+def plot_transformed(x_f, y_f):
     plt.plot(x_f, np.abs(y_f))
     plt.show()
 
@@ -130,7 +139,7 @@ handler = Handler.CSV_handler()
 file = "/Exp20201205_2myo_hardTypePP/HaluskaMarek_20201207_1810/myoLeftEmg.csv"
 df = handler.get_time_emg_table(file, 1)
 #plot_df(df)
-trans_df = DataFrame(transformed_df(df))
+x_f, y_f = denoise_signal_pywt(df)
 #print(trans_df.info)
-plot_fft(trans_df)
+plot_transformed(x_f, y_f)
 #'''