Update app.py

app.py

@@ -35,13 +35,53 @@ import traceback
 # ====================== Hugging Face Tesseract 环境 ========================== 
 import pytesseract
 import os
+import shutil
 
-# 硬编码 Hugging Face 环境的 Tesseract 路径
-pytesseract.pytesseract.tesseract_cmd = r'/user/bin/tesseract'
+# 根据环境自动检测 Tesseract 路径
+def setup_tesseract():
+    # 尝试所有可能的 Tesseract 路径
+    possible_paths = [
+        '/usr/bin/tesseract',      # Hugging Face Spaces 或 Linux 标准路径
+        '/usr/local/bin/tesseract', # Linux/macOS 本地安装
+        '/opt/homebrew/bin/tesseract', # macOS Homebrew
+        '/opt/tesseract/bin/tesseract', # 某些 Linux 发行版
+        'tesseract',               # 系统PATH中
+    ]
+    
+    # 首先尝试使用 which 命令查找 tesseract
+    tesseract_path = shutil.which('tesseract')
+    if tesseract_path:
+        pytesseract.pytesseract.tesseract_cmd = tesseract_path
+        print(f"Found Tesseract at: {tesseract_path}")
+        return
+    
+    # 如果 which 找不到，尝试常见路径
+    for path in possible_paths:
+        if path != 'tesseract' and os.path.exists(path):
+            pytesseract.pytesseract.tesseract_cmd = path
+            print(f"Found Tesseract at: {path}")
+            return
+    
+    # 如果是 'tesseract' 字符串，直接设置（让系统在PATH中查找）
+    if 'tesseract' in possible_paths:
+        pytesseract.pytesseract.tesseract_cmd = 'tesseract'
+        print("Using 'tesseract' from system PATH")
+        return
+    
+    # 如果都找不到，提供详细的错误信息
+    print("Tesseract not found in any of the expected locations:")
+    for path in possible_paths:
+        print(f"  - {path}")
+    print("\nTroubleshooting steps:")
+    print("1. Ensure packages.txt contains 'tesseract-ocr'")
+    print("2. Try rebuilding the Space")
+    print("3. Check if the base image supports apt packages")
+    
+    # 不抛出错误，而是设置一个默认值，让 pytesseract 自己处理
+    pytesseract.pytesseract.tesseract_cmd = 'tesseract'
 
-# 验证路径是否存在（调试用）
-if not os.path.exists(pytesseract.pytesseract.tesseract_cmd):
-    raise RuntimeError("Tesseract not found! Check packages.txt and rebuild.")
+# 设置 Tesseract
+setup_tesseract()
 
 # ====================== Module 1: PDF/Image to Markdown ======================
 class PDFImageToMarkdown: