import numpy as np
import os


AU = ["Person", "Case", "Confidence", "AU01", "AU02", "AU04", "AU05", "AU06", "AU07", "AU09","AU10","AU12","AU14","AU15","AU17","AU20","AU23",
  "AU25","AU26", "AU45"]

# folder where output csv files of OpenFace are stored
folder = '/home/valapil/Project/ForkCausal_Adithya/csv_files'
csv_files = [file for file in os.listdir(folder) if file.endswith('.csv')]

# load the offset values
offset = np.load(f'/home/valapil/Project/ForkCausal_Adithya/offset_values.npy')
length = {}
same = np.unique(offset[1:, 0])
for i in same:
    same_rows = offset[offset[:, 0] == i]
    length[i] = max(abs(int(same_rows[0][2]) - int(same_rows[0][1])), abs(int(same_rows[1][2]) - int(same_rows[1][1])))

for file in csv_files:
    base_name, extension = os.path.splitext(file)
    new_name = base_name.split('.')[0]
    f1 = new_name[:-2]
    f2 = new_name[-1:]
    file_path = os.path.join(folder, file)
    len = length[new_name[:-3] + new_name[-1]]
    start = np.where(offset[1:, 3]  == new_name)[0][0]
    start = int(offset[start+1, 1])
    data_all = np.genfromtxt(file_path, delimiter=',', skip_header=1)
    data = data_all[start:len-200, 679:696]

    selected_columns = np.hstack((np.full((data.shape[0], 2), [f1, f2]), data_all[start:len-200, [3]], data))

    AU = np.row_stack((AU, selected_columns))

    #  save each offset data here
    np.save(f'/home/valapil/Project/ForkCausal_Adithya/data/{new_name}.npy', data)

#  save all offset data with the particular case and confidence value here
np.save(f'/home/valapil/Project/ForkCausal_Adithya/processed_data.npy', AU)