Filter-Similar-Files.py

from tkinter import Tcl #OK
import os #OK
import numpy as np #OK
import cv2 #OK
import time

# Locatia standard pentru toate frame-urile
folder_frames = '/Users/cristianpogan/Downloads/Test-comparatie-frames/outputlit2/Test-Data-24-Apr/No-Bin-24-Apr/'
extension = ".jpeg"

# Numara cate fisiere sunt in folder
path, dirs, files = next(
    os.walk(folder_frames))
file_count = len(files)
print("Sunt ", file_count, " frame-uri")

print("-------------------")

file_list = os.listdir(folder_frames)
# Face un array cu denuminile fisierelor .jpeg
lista_frameuri = list(Tcl().call('lsort', '-dict', file_list))

#dir = sorted(os.listdir(os.getcwd()), key=len)
#print(lista_frameuri)

start = time.perf_counter()
over_threshold = 0


for x in range(len(lista_frameuri)-1):

    print(lista_frameuri[x])
    if lista_frameuri[x].endswith(extension):
        # ----------HASH comparison------------------------------------------
        frameX1 = os.path.join(folder_frames, lista_frameuri[x])
        frameX2 = os.path.join(folder_frames, lista_frameuri[x+1])
        # ----------Pregatirea imaginilor pentru MSE si SSIM----------------
        # load the images -- the original, the original + contrast,
        # and the original + photoshop
        frameX1CV = cv2.imread(frameX1)
        frameX2CV = cv2.imread(frameX2)
        # convert the images to grayscale
        image1 = cv2.cvtColor(frameX1CV, cv2.COLOR_BGR2GRAY)
        image2 = cv2.cvtColor(frameX2CV, cv2.COLOR_BGR2GRAY)
        # ----------MSE - Mean Squared Error - comparison-------------------
        MSEerr_result = np.sum((image1.astype("float") - image2.astype("float")) ** 2)
        MSEerr_result /= float(image1.shape[0] * image1.shape[1])
        # ----------Afisarea rezultatelor----------------------------------
        print(x, " intre: ", frameX1[-42:-5], "si", frameX2[-42:-5] , ": ", int(MSEerr_result))

        if MSEerr_result < 30:
            over_threshold = over_threshold + 1
            print("Removed ", frameX1)
            os.remove(frameX1)
        # Stergerea fisierelor asemanatoare
        #if diff == 0 or diff == 1:
        #    print("Removed ", frameX1)
        #    os.remove(frameX1)
    else:
        pass

print("over_threshold: ", over_threshold)
end = time.perf_counter()
print(end - start)