integrate in one UI

This commit is contained in:
ccppoo 2020-01-03 01:53:57 +09:00
parent 7303c3d8ec
commit f5a1b34066
2 changed files with 137 additions and 52 deletions

View File

@ -29,8 +29,8 @@ except ImportError as e:
print("install Anaconda : https://www.anaconda.com/distribution/ \n")
class Decensor(QtCore.QThread):
def __init__(self, text_edit = None, text_cursor = None, ui_mode = None):
super().__init__()
def __init__(self, parentThread = None, text_edit = None, text_cursor = None, ui_mode = None):
super().__init__(parentThread)
args = config.get_args()
self.is_mosaic = args.is_mosaic
self.variations = args.variations
@ -41,20 +41,29 @@ class Decensor(QtCore.QThread):
self.signals = None # Signals class will be given by progressWindow
if ui_mode is not None:
self.ui_mode = ui_mode
else:
self.ui_mode = args.ui_mode
self.model = None
self.warm_up = False
# if ui_mode is not None:
# self.ui_mode = ui_mode
# else:
# self.ui_mode = args.ui_mode
#
# if self.ui_mode:
# self.text_edit = text_edit
# self.text_cursor = text_cursor
# self.ui_mode = True
if not os.path.exists(self.decensor_output_path):
os.makedirs(self.decensor_output_path)
if self.ui_mode:
self.text_edit = text_edit
self.text_cursor = text_cursor
self.ui_mode = True
def run(self):
if not self.warm_up :
print("if self.warm_up :")
self.load_model()
return
elif self.warm_up:
print("elif not self.warm_up:")
self.decensor_all_images_in_folder()
def stop(self):
@ -69,15 +78,23 @@ class Decensor(QtCore.QThread):
return mask
def load_model(self):
self.signals.update_progress_LABEL.emit("load_model()", "Loading neural network. This may take a while.")
self.signals.insertText_progressCursor.emit("Loading model ... please wait ...\n")
if self.model is None :
self.model = InpaintNN(bar_model_name = "./models/bar/Train_775000.meta",
bar_checkpoint_name = "./models/bar/",
mosaic_model_name = "./models/mosaic/Train_290000.meta",
mosaic_checkpoint_name = "./models/mosaic/",
is_mosaic=self.is_mosaic)
self.warm_up = True
print("load model finished")
self.signals.insertText_progressCursor.emit("Loading model finished!\n")
self.signals.update_decensorButton_Text.emit("Decensor Your Images")
self.signals.update_decensorButton_Enabled.emit(True)
def decensor_all_images_in_folder(self):
#load model once at beginning and reuse same model
if not self.warm_up :
# incase of running by source code
self.load_model()
input_color_dir = self.decensor_input_path
@ -87,24 +104,37 @@ class Decensor(QtCore.QThread):
output_dir = self.decensor_output_path
# Change False to True before release --> file.check_file(input_dir, output_dir, True)
self.signals.update_progress_LABEL.emit("file.check_file()", "Checking image files and directory...")
# self.signals.update_progress_LABEL.emit("file.check_file()", "Checking image files and directory...")
self.signals.insertText_progressCursor.emit("Checking image files and directory...\n")
file_names, self.files_removed = file.check_file(input_dir, output_dir, False)
self.signals.total_ProgressBar_update_MAX_VALUE.emit("set total progress bar MaxValue : "+str(len(file_names)),len(file_names))
# self.signals.total_ProgressBar_update_MAX_VALUE.emit("set total progress bar MaxValue : "+str(len(file_names)),len(file_names))
'''
print("set total progress bar MaxValue : "+str(len(file_names)))
self.signals.update_ProgressBar_MAX_VALUE.emit(len(file_names))
'''
self.signals.insertText_progressCursor.emit("Decensoring {} image files\n".format(len(file_names)))
#convert all images into np arrays and put them in a list
for n, file_name in enumerate(file_names, start = 1):
self.signals.total_ProgressBar_update_VALUE.emit("Decensoring {} / {}".format(n, len(file_names)), n)
# self.signals.total_ProgressBar_update_VALUE.emit("Decensoring {} / {}".format(n, len(file_names)), n)
'''
self.update_ProgressBar_SET_VALUE.emit(n)
print("Decensoring {} / {}".format(n, len(file_names)))
'''
self.signals.insertText_progressCursor.emit("Decensoring image file : {}\n".format(file_name))
# signal progress bar value == masks decensored on image ,
# e.g) sample image : 17
self.signals.signal_ProgressBar_update_VALUE.emit("reset value", 0) # set to 0 for every image at start
self.signals.update_progress_LABEL.emit("for-loop, \"for file_name in file_names:\"","Decensoring : "+str(file_name))
# self.signals.signal_ProgressBar_update_VALUE.emit("reset value", 0) # set to 0 for every image at start
# self.signals.update_progress_LABEL.emit("for-loop, \"for file_name in file_names:\"","Decensoring : "+str(file_name))
color_file_path = os.path.join(input_color_dir, file_name)
color_basename, color_ext = os.path.splitext(file_name)
if os.path.isfile(color_file_path) and color_ext.casefold() == ".png":
print("--------------------------------------------------------------------------")
print("Decensoring the image {}".format(color_file_path))
print("Decensoring the image {}\n".format(color_file_path))
try :
colored_img = Image.open(color_file_path)
except:
@ -131,20 +161,25 @@ class Decensor(QtCore.QThread):
else: #for...else, i.e if the loop finished without encountering break
print("Corresponding original, uncolored image not found in {}".format(color_file_path))
print("Check if it exists and is in the PNG or JPG format.")
self.signals.insertText_progressCursor.emit("Corresponding original, uncolored image not found in {}\n".format(color_file_path))
self.signals.insertText_progressCursor.emit("Check if it exists and is in the PNG or JPG format.\n")
#if we are doing a bar decensor
else:
self.decensor_image_variations(colored_img, colored_img, file_name)
else:
print("--------------------------------------------------------------------------")
print("Image can't be found: "+str(color_file_path))
self.signals.insertText_progressCursor.emit("Image can't be found: "+str(color_file_path) + "\n")
print("--------------------------------------------------------------------------")
if self.files_removed is not None:
file.error_messages(None, self.files_removed)
print("\nDecensoring complete!")
#unload model to prevent memory issues
self.signals.update_progress_LABEL.emit("finished", "Decensoring complete! Close this window and reopen DCP to start a new session.")
# self.signals.update_progress_LABEL.emit("finished", "Decensoring complete! Close this window and reopen DCP to start a new session.")
self.signals.insertText_progressCursor.emit("\nDecensoring complete! remove decensored file before decensoring again not to overwrite")
self.signals.update_decensorButton_Enabled.emit(True)
tf.reset_default_graph()
def decensor_image_variations(self, ori, colored, file_name=None):
@ -197,16 +232,25 @@ class Decensor(QtCore.QThread):
#colored image is only used for finding the regions
regions = find_regions(colored.convert('RGB'), [v*255 for v in self.mask_color])
print("Found {region_count} censored regions in this image!".format(region_count = len(regions)))
self.signals.insertText_progressCursor.emit("Found {region_count} censored regions in this image!".format(region_count = len(regions)))
if len(regions) == 0 and not self.is_mosaic:
print("No green regions detected! Make sure you're using exactly the right color.")
print("No green (0,255,0) regions detected! Make sure you're using exactly the right color.")
self.signals.insertText_progressCursor.emit("No green (0,255,0) regions detected! Make sure you're using exactly the right color.\n")
return
self.signals.signal_ProgressBar_update_MAX_VALUE.emit("Found {} masked regions".format(len(regions)), len(regions))
# self.signals.signal_ProgressBar_update_MAX_VALUE.emit("Found {} masked regions".format(len(regions)), len(regions))
print("Found {} masked regions".format(len(regions)))
# self.signals.insertText_progressCursor.emit("Found {} masked regions\n".format(len(regions)))
self.signals.update_ProgressBar_MAX_VALUE.emit(len(regions))
self.signals.update_ProgressBar_SET_VALUE.emit(0)
output_img_array = ori_array[0].copy()
for region_counter, region in enumerate(regions, 1):
self.signals.update_progress_LABEL.emit("\"Decensoring regions in image\"","Decensoring censor {}/{}".format(region_counter,len(regions)))
# self.signals.update_progress_LABEL.emit("\"Decensoring regions in image\"","Decensoring censor {}/{}".format(region_counter,len(regions)))
self.signals.insertText_progressCursor.emit("Decensoring regions in image, Decensoring censor {}/{}".format(region_counter,len(regions)))
bounding_box = expand_bounding(ori, region, expand_factor=1.5)
crop_img = ori.crop(bounding_box)
# crop_img.show()
@ -286,7 +330,9 @@ class Decensor(QtCore.QThread):
bounding_height_index = row + bounding_box[1]
if (bounding_width_index, bounding_height_index) in region:
output_img_array[bounding_height_index][bounding_width_index] = pred_img_array[i,:,:,:][row][col]
self.signals.signal_ProgressBar_update_VALUE.emit("{} out of {} regions decensored.".format(region_counter, len(regions)), region_counter)
# self.signals.signal_ProgressBar_update_VALUE.emit("{} out of {} regions decensored.".format(region_counter, len(regions)), region_counter)
self.signals.update_ProgressBar_SET_VALUE.emit(region_counter)
self.signals.insertText_progressCursor.emit("{} out of {} regions decensored.\n".format(region_counter, len(regions)))
print("{region_counter} out of {region_count} regions decensored.".format(region_counter=region_counter, region_count=len(regions)))
output_img_array = output_img_array * 255.0
@ -298,7 +344,9 @@ class Decensor(QtCore.QThread):
output_img = Image.fromarray(output_img_array.astype('uint8'))
output_img = self.apply_variant(output_img, variant_number)
self.signals.update_progress_LABEL.emit("current image finished", "Decensoring of current image finished. Saving image...")
# self.signals.update_progress_LABEL.emit("current image finished", "Decensoring of current image finished. Saving image...")
self.signals.insertText_progressCursor.emit("Decensoring of current image finished. Saving image...")
print("current image finished")
if file_name != None:
#save the decensored image
@ -306,10 +354,11 @@ class Decensor(QtCore.QThread):
file_name = base_name + " " + str(variant_number) + ext
save_path = os.path.join(self.decensor_output_path, file_name)
output_img.save(save_path)
print("Decensored image saved to {save_path}!".format(save_path=save_path))
return
self.signals.insertText_progressCursor.emit("Decensored image saved to {save_path}!".format(save_path=save_path))
self.signals.insertText_progressCursor.emit("="*30)
else:
# Legacy Code piece ↓, used when DCPv1 had ui with Painting
print("Decensored image. Returning it.")
return output_img

72
main.py
View File

@ -5,19 +5,27 @@
# The greater the number of variations, the longer decensoring process will be.
import sys, time
from PySide2.QtWidgets import QWidget, QHBoxLayout, QVBoxLayout, QGridLayout, QGroupBox, QDesktopWidget, QApplication, QAction, qApp, QApplication, QMessageBox, QRadioButton, QPushButton, QTextEdit, QLabel, QSizePolicy,QMainWindow
from PySide2.QtWidgets import QWidget, QHBoxLayout, QVBoxLayout, QGridLayout, QGroupBox, QDesktopWidget, QApplication
from PySide2.QtWidgets import QAction, qApp, QApplication, QMessageBox, QRadioButton, QPushButton, QTextEdit, QLabel
from PySide2.QtWidgets import QSizePolicy,QMainWindow, QStatusBar, QProgressBar
from PySide2.QtCore import Qt, QObject
from PySide2.QtGui import QFont, QTextCursor
from decensor import Decensor
from signals import Signals
from progressWindow import ProgressWindow
# from decensor import Decensor
# from progressWindow import ProgressWindow
class MainWindow(QWidget):
def __init__(self):
super().__init__()
self.signals = Signals()
self.initUI()
self.setSignals()
self.decensor = Decensor(self)
self.load_model()
def initUI(self):
@ -75,25 +83,60 @@ class MainWindow(QWidget):
self.progressMessage.setReadOnly(True)
self.progressCursor.insertText("After you prepared your images, click on the decensor button once to begin decensoring.\nPlease be patient.\nDecensoring will take time.\n")
# Progress Bar
self.statusBar = QStatusBar(self)
self.progressBar = QProgressBar()
self.progressBar.setMinimum(0)
self.progressBar.setMaximum(100)
self.progressBar.setValue(0)
self.statusLabel = QLabel("Showing Progress")
self.statusBar.addWidget(self.statusLabel, 1)
self.statusBar.addWidget(self.progressBar, 2)
#put all groups into grid
# addWidget(row, column, rowSpan, columnSpan)
grid_layout.addWidget(self.tutorialLabel, 0, 0, 1, 2)
grid_layout.addWidget(self.censorTypeGroupBox, 1, 0, 1, 1)
grid_layout.addWidget(self.variationsGroupBox, 1, 1, 1, 1)
grid_layout.addWidget(self.decensorButton, 2, 0, 1, 2)
grid_layout.addWidget(self.progressMessage, 3, 0, 4, 2)
grid_layout.addWidget(self.progressMessage, 3, 0, 1, 2)
grid_layout.addWidget(self.statusBar, 4, 0, 1, 2)
#window size settings
self.resize(500, 500)
self.resize(900, 600)
self.center()
self.setWindowTitle('DeepCreamPy v2.2.0-beta')
self.show()
def load_model(self):
# load model to make able to decensor several times
self.decensorButton.setEnabled(False)
self.decensorButton.setText("Loading Machine Learning Model (Please Wait...)")
self.decensor.start()
self.decensor.signals = self.signals
self.progressCursor.insertText("Loading Decensor app consumes 6 GB memory at maximum")
def setSignals(self):
self.signals.update_decensorButton_Text.connect(self.decensorButton.setText)
self.signals.update_decensorButton_Enabled.connect(self.decensorButton.setEnabled)
self.signals.update_statusLabel_Text.connect(self.statusLabel.setText)
self.signals.update_ProgressBar_SET_VALUE.connect(self.progressBar.setValue)
self.signals.update_ProgressBar_MAX_VALUE.connect(self.progressBar.setMaximum)
self.signals.update_ProgressBar_MIN_VALUE.connect(self.progressBar.setMinimum)
# self.signals.insertText_progressCursor.connect(self.progressCursor.insertText)
self.signals.insertText_progressCursor.connect(self.progressMessage.append)
self.signals.clear_progressMessage.connect(self.progressMessage.clear)
self.signals.appendText_progressMessage.connect(self.progressMessage.append)
def decensorClicked(self):
self.decensorButton.setEnabled(False)
self.progressMessage.clear()
self.progressCursor.insertText("Decensoring has begun!\n")
decensor = Decensor(text_edit = self.progressMessage, text_cursor = self.progressCursor, ui_mode = True)
# for now, decensor is initiated when this app is started
# self.decensor = Decensor(text_edit = self.progressMessage, text_cursor = self.progressCursor, ui_mode = True)
#https://stackoverflow.com/questions/42349470/pyqt-find-checked-radiobutton-in-a-group
#set decensor to right settings
#censor type
@ -103,9 +146,9 @@ class MainWindow(QWidget):
if cb.isChecked():
censorType = cb.text()
if censorType == 'Bar censor':
decensor.is_mosaic = False
self.decensor.is_mosaic = False
else:
decensor.is_mosaic = True
self.decensor.is_mosaic = True
#variations count
variationsElements = self.variationsGroupBox.children()
@ -113,20 +156,13 @@ class MainWindow(QWidget):
for vb in variationsButtons:
if vb.isChecked():
variations = int(vb.text())
decensor.variations = variations
self.decensor.variations = variations
self.decensorButton.setEnabled(True)
self.hide()
self.progress = ProgressWindow(self, decensor = decensor)
self.decensorButton.setEnabled(False)
self.decensor.start()
# decensor.decensor_all_images_in_folder()
# self.progress.hide()
# self.show()
# def showAbout(self):
# QMessageBox.about(self, 'About', "DeepCreamPy v2.2.0 \n Developed by deeppomf")
# #centers the main window
def center(self):
qr = self.frameGeometry()