Add a script to find unused images

d-run · Dec 24, 2024 · 4850e1e · 4850e1e
1 parent 6eb9437
commit 4850e1e
Show file tree

Hide file tree

Showing 12 changed files with 31 additions and 0 deletions.
diff --git a/docs/zh/docs/dmc/images/create-model.png b/docs/zh/docs/dmc/images/create-model.png
diff --git a/docs/zh/docs/dmc/images/dmc-home.png b/docs/zh/docs/dmc/images/dmc-home.png
diff --git a/docs/zh/docs/dmc/images/inner01.png b/docs/zh/docs/dmc/images/inner01.png
diff --git a/docs/zh/docs/dmc/images/inner02.png b/docs/zh/docs/dmc/images/inner02.png
diff --git a/docs/zh/docs/dmc/images/inner03.jpg b/docs/zh/docs/dmc/images/inner03.jpg
diff --git a/docs/zh/docs/dmc/images/inner03.png b/docs/zh/docs/dmc/images/inner03.png
diff --git a/docs/zh/docs/dmc/images/local01.png b/docs/zh/docs/dmc/images/local01.png
diff --git a/docs/zh/docs/dmc/images/local04.png b/docs/zh/docs/dmc/images/local04.png
diff --git a/docs/zh/docs/dmc/images/online04.png b/docs/zh/docs/dmc/images/online04.png
diff --git a/docs/zh/docs/dmc/images/tuner01.png b/docs/zh/docs/dmc/images/tuner01.png
diff --git a/docs/zh/docs/dmc/model-service/local.md b/docs/zh/docs/dmc/model-service/local.md
@@ -19,6 +19,7 @@ hide:
 1. 点击模型服务右边的 ┇ 符号，选择 **绑定/解绑工作空间** ，即可将模型服务分享分享到其他工作空间下。绑定后，其他工作空间下的用户就可以直接使用该模型服务。
 
     ![绑定/解绑工作空间1](../images/local03.jpg)
+
     ![绑定/解绑工作空间2](../images/local04.jpg)
 
 2. 去到分享的工作空间下可查看分享的模型服务，并可以直接使用该模型服务进行 **对话调式** 和 **应用创建** 。

diff --git a/scripts/find_unused_images.py b/scripts/find_unused_images.py
@@ -0,0 +1,30 @@
+import os
+import re
+
+# 图片目录
+image_dir = 'docs/zh/docs/dmc/images'  # 确保路径正确
+# Markdown 文件目录
+markdown_dir = 'docs/zh/docs/dmc'
+
+# 收集所有图片文件
+images = {f for f in os.listdir(image_dir) if f.endswith(('.jpg', '.png'))}
+
+# 收集在 Markdown 文件中引用的图片
+referenced_images = set()
+
+# 使用 os.walk 遍历 Markdown 文件目录及其所有子文件夹
+for root, _, files in os.walk(markdown_dir):
+    for file in files:
+        if file.endswith('.md'):
+            with open(os.path.join(root, file), 'r', encoding='utf-8') as f:
+                content = f.read()
+                # 使用正则表达式查找所有可能的图片引用
+                # 修改正则表达式以匹配 "../images/" 前缀
+                matches = re.findall(r'!\[.*?\]\(\.\./images/(.*?\.(?:jpg|png))\)', content)
+                # 提取文件名部分并添加到 referenced_images
+                referenced_images.update(matches)
+
+# 找到未被引用的图片
+unused_images = images - referenced_images
+
+print("未被引用的图片文件:", unused_images)