Update app2.py

app2.py

@@ -166,10 +166,10 @@ class URLProcessor:
 class FileProcessor:
     """Class to handle file processing"""
     
-    def __init__(self, max_file_size: int = 10 * 1024 * 1024):  # 10MB default
+    def __init__(self, max_file_size: int = 2 * 1024 * 1024 * 1024):  # 2GB default
         self.max_file_size = max_file_size
         self.supported_text_extensions = {'.txt', '.md', '.csv', '.json', '.xml'}
-
+        
     def is_text_file(self, filepath: str) -> bool:
         """Check if file is a text file"""
         try:
@@ -226,25 +226,41 @@ class FileProcessor:
                             logger.error(f"Error reading file {filename}: {str(e)}")
         return results
 
-    def _process_single_file(self, file) -> List[Dict]:
-        try:
-            file_stat = os.stat(file.name)
+def _process_single_file(self, file) -> List[Dict]:
+    try:
+        file_stat = os.stat(file.name)
+        
+        # For very large files, read in chunks and summarize
+        if file_stat.st_size > 100 * 1024 * 1024:  # 100MB
+            logger.info(f"Processing large file: {file.name} ({file_stat.st_size} bytes)")
+            
+            # Read first and last 1MB for extremely large files
+            content = ""
+            with open(file.name, 'r', encoding='utf-8', errors='ignore') as f:
+                content = f.read(1 * 1024 * 1024)  # First 1MB
+                content += "\n...[Content truncated due to large file size]...\n"
+                
+                # Seek to the last 1MB
+                f.seek(max(0, file_stat.st_size - 1 * 1024 * 1024))
+                content += f.read()  # Last 1MB
+        else:
+            # Regular file processing
             with open(file.name, 'r', encoding='utf-8', errors='ignore') as f:
                 content = f.read()
-            
-            return [{
-                'source': 'file',
-                'filename': os.path.basename(file.name),
-                'file_size': file_stat.st_size,
-                'mime_type': mimetypes.guess_type(file.name)[0],
-                'created': datetime.fromtimestamp(file_stat.st_ctime).isoformat(),
-                'modified': datetime.fromtimestamp(file_stat.st_mtime).isoformat(),
-                'content': content,
-                'timestamp': datetime.now().isoformat()
-            }]
-        except Exception as e:
-            logger.error(f"File processing error: {e}")
-            return []
+        
+        return [{
+            'source': 'file',
+            'filename': os.path.basename(file.name),
+            'file_size': file_stat.st_size,
+            'mime_type': mimetypes.guess_type(file.name)[0],
+            'created': datetime.fromtimestamp(file_stat.st_ctime).isoformat(),
+            'modified': datetime.fromtimestamp(file_stat.st_mtime).isoformat(),
+            'content': content,
+            'timestamp': datetime.now().isoformat()
+        }]
+    except Exception as e:
+        logger.error(f"File processing error: {e}")
+        return []
 
 def create_interface():
     """Create a comprehensive Gradio interface with advanced features"""