GenerateTxT.配合danbooruDownloader图片抓取教程使用.py

#!/usr/bin/python
print("me")
import os
import xml
import xml.sax
import sys
img_folder = "F:\StableDifussion\stable diffusion 工具\DanbooruDownloader3.20220828\images\gelbooru.com\deedlit\proccessed"
xml_path = "F:\StableDifussion\stable diffusion 工具\DanbooruDownloader3.20220828\savelist\deedlit.xml"

xml_imgs_list = [];


class MovieHandler( xml.sax.ContentHandler ):
   def __init__(self):
      self.CurrentData = ""
      self.Id = ""
      self.index = 0
      self.Filename = ""
      self.Tags = ""
      self.row = []
 
   # 元素开始事件处理
   def startElement(self, tag, attributes):
      self.CurrentData = tag
      #print('self.CurrentData:'+str(self.CurrentData))
		 
   # 元素结束事件处理
   def endElement(self, tag):
      if self.CurrentData == "JpegHeight":
        #xml_imgs_list.append(self.row) 
        #self.row = []
        #self.index = self.index + 1
        print ("JpegHeight:")
        #print (self.type)
      self.CurrentData = ""
   # 内容事件处理
   def characters(self, content):
      if self.CurrentData == "Id":
         print('Id:'+str(content))
         #print('self.Id:'+str(content))
         #self.Id =content
         #print('xml_imgs_list:')
         #print(xml_imgs_list)
         self.row = []
         self.row.append(content)
      elif self.CurrentData == "Tags":
         print('Tags:'+str(content))
         #self.Tags = content
         self.row.append(content)
      elif self.CurrentData == "Filename":
         print('Filename:'+str(content))
         #print('self.Id:'+str(self.Id)+'\n')
         #print('self.index:'+str(self.index)+'\n')
         #self.Filename = content
         self.row.append(content) 
         if len(self.row) < 3:
            print(self.row)
            print("该图片生成失败,按回车键继续")
            input()
            self.row = []
         else:
            xml_imgs_list.append(self.row) 
            self.row = []
      
parser  = xml.sax.make_parser()
parser.setFeature(xml.sax.handler.feature_namespaces, 0)
Handler = MovieHandler()
parser.setContentHandler( Handler )
parser.parse(xml_path);



prompt_arr = []
for file_name in os.listdir(img_folder): 
	for xml_row in xml_imgs_list:
		try:
			if file_name in xml_row[2]:
				text_content = xml_row[1]
				prompt_arr.append(text_content)
				file_path_explode_arr = os.path.splitext(file_name);
				#save_path = img_folder+file_path_explode_arr[0]+'.txt'
				save_path = os.path.join(img_folder,file_path_explode_arr[0]+".txt")
				##print("save_path:"+save_path)
				##print("\n\r")
				##print("text_content:"+text_content)
				##print("\n\r")
				#fwrite(save_path,text_content);
				
				file = open(save_path,'w+')
				file.write(text_content);
				file.close()
		except:
			print(xml_row)
			print("该图片生成失败,按回车键继续")
			input()
			#print("OS error: {0}".format(err))
file = open(os.path.join(img_folder,"all_prompt.txt"),'w+')
file.write(",".join(prompt_arr));
file.close()