Update app.py

app.py

@@ -8,6 +8,22 @@ from PIL import Image
 import fitz  # PyMuPDF
 import numpy as np
 from transformers import NougatProcessor, VisionEncoderDecoderModel
+import nltk
+import ssl
+
+# 初始化NLTK
+try:
+    _create_unverified_https_context = ssl._create_unverified_context
+except AttributeError:
+    pass
+else:
+    ssl._create_default_https_context = _create_unverified_https_context
+
+# 下载NLTK必要的数据
+try:
+    nltk.data.find('tokenizers/punkt')
+except LookupError:
+    nltk.download('punkt')
 
 # Set environment variables
 HF_TOKEN = os.environ.get("HF_TOKEN", None)
@@ -90,6 +106,20 @@ def extract_text_from_pdf(pdf_bytes):
         return default_paper_content
     
     try:
+        # 确保NLTK已安装
+        try:
+            import nltk
+            try:
+                nltk.data.find('tokenizers/punkt')
+            except LookupError:
+                nltk.download('punkt')
+        except ImportError:
+            print("Installing NLTK...")
+            import subprocess
+            subprocess.check_call(["pip", "install", "nltk", "python-Levenshtein"])
+            import nltk
+            nltk.download('punkt')
+        
         # Load Nougat model
         processor, model = load_nougat_model()