add json

count_media_types_and_check_files.py

@@ -0,0 +1,71 @@
+import json
+import os
+from tqdm import tqdm  # 导入 tqdm 库
+
+def count_media_types_and_check_files(json_file, base_paths):
+    counts = {
+        "audio_caption": {"total": 0, "missing": 0},
+        "audio_asr": {"total": 0, "missing": 0},
+        "image": {"total": 0, "missing": 0},
+        "video": {"total": 0, "missing": 0}
+    }
+    audio_caption_items = []
+    with open(json_file, 'r', encoding='utf-8') as file:
+        data = json.load(file)
+        # 使用 tqdm 包装循环，显示进度条
+        for item in tqdm(data, desc="Processing files", unit="file"):
+            if "audio_caption" in item:
+                counts["audio_caption"]["total"] += 1
+                file_path = os.path.join(base_paths["audio_caption"], item["audio_caption"])
+                if not os.path.exists(file_path):
+                    counts["audio_caption"]["missing"] += 1
+                if item["audio_caption"].startswith("vggsound"):
+                    continue
+                audio_caption_items.append(item)
+            
+            if "image" in item:
+                counts["image"]["total"] += 1
+                file_path = os.path.join(base_paths["image"], item["image"])
+                if not os.path.exists(file_path):
+                    counts["image"]["missing"] += 1
+                audio_caption_items.append(item)
+            '''
+            if "video" in item:
+                counts["video"]["total"] += 1
+                file_path = os.path.join(base_paths["video"], item["video"])
+                if not os.path.exists(file_path):
+                    counts["video"]["missing"] += 1
+                else:
+                 #   audio_caption_items.append(item)
+            '''
+            if "audio_asr" in item:
+                counts["audio_asr"]["total"] += 1
+                file_path = os.path.join(base_paths["audio_asr"], item["audio_asr"])
+                if not os.path.exists(file_path):
+                    counts["audio_asr"]["missing"] += 1
+                audio_caption_items.append(item)
+            
+
+    '''
+    # 重新编号 id
+    for i, item in enumerate(audio_caption_items):
+        item["id"] = str(i)
+    audio_caption_output_file = 'image_stage3_04013.json'
+    # 写入audio_caption的 JSON 文件
+    with open(audio_caption_output_file, 'w', encoding='utf-8') as file:
+        json.dump(audio_caption_items, file, ensure_ascii=False, indent=4)
+    '''
+    return counts
+
+
+
+# 使用示例
+json_file = 'image_asr_caption_stage3_0409.json'
+base_paths = {
+    "audio_caption": "/media/yu/18TB/audio_asr",
+    "audio_asr": "/home/yu/Downloads",
+    "image": "/media/yu/18TB/videollava",
+    "video": "/media/yu/18TB/videollava"
+}
+counts = count_media_types_and_check_files(json_file, base_paths)
+print(counts)

image_asr_caption_stage3_0409.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03bb48530a9c1e3654b970b56312c2c8f8ae7222cb15a34148d07412b52e1fb3
+size 1496205036