Update app.py

app.py

@@ -33,6 +33,65 @@ for index, row in df.iterrows():
         continue
 
 print("所有行处理完成")
+
+
+import pandas as pd
+import pathlib
+import numpy as np
+import shutil
+import os
+
+# 获取HunyuanVideo-Foley目录下的MP4文件路径
+l0 = pd.Series(list(pathlib.Path(".").rglob("*.mp4"))).map(str).map(
+    lambda x: x if "17563" in x else np.nan
+).dropna().sort_values().values.tolist()
+
+# 读取CSV文件
+df = pd.read_csv("wan_gen_videos_captioned/metadata.csv")
+print(df.shape, len(l0))
+df["file_name"] = pd.Series(l0).map(lambda x: x.split("/")[-1]).values.tolist()
+
+# 创建目标文件夹
+target_folder = "wan_gen_videos_HunyuanVideo_Foley_sound_captioned"
+os.makedirs(target_folder, exist_ok=True)
+
+# 处理每一行数据
+for index, row in df.iterrows():
+    try:
+        # 获取源文件路径和目标文件路径
+        source_file = None
+        for file_path in l0:
+            if row['file_name'] in file_path:
+                source_file = file_path
+                break
+        
+        if source_file:
+            # 复制MP4文件
+            target_mp4 = os.path.join(target_folder, row['file_name'])
+            shutil.copy2(source_file, target_mp4)
+            
+            # 创建对应的文本文件
+            txt_filename = os.path.splitext(row['file_name'])[0] + '.txt'
+            txt_filepath = os.path.join(target_folder, txt_filename)
+            
+            with open(txt_filepath, 'w', encoding='utf-8') as f:
+                f.write(row['prompt'])
+            
+            print(f"成功处理: {row['file_name']} -> {txt_filename}")
+        else:
+            print(f"警告: 未找到文件 {row['file_name']}")
+            
+    except Exception as e:
+        print(f"处理失败 - 行 {index}: {str(e)}")
+        continue
+
+print("所有文件处理完成！")
+print(f"文件已保存到: {target_folder}")
+
+df.to_csv("wan_gen_videos_HunyuanVideo_Foley_sound_captioned/metadata.csv", index = False)
+
+!cp README.md wan_gen_videos_HunyuanVideo_Foley_sound_captioned
+!huggingface-cli upload svjack/wan_gen_videos_HunyuanVideo_Foley_sound_captioned wan_gen_videos_HunyuanVideo_Foley_sound_captioned --repo-type dataset
 '''
 
 import spaces