mirror of
				https://github.com/kevinveenbirkenbach/directory-content-scanner.git
				synced 2025-11-04 03:48:22 +00:00 
			
		
		
		
	implemented string whitelist
This commit is contained in:
		
							
								
								
									
										10
									
								
								README.md
									
									
									
									
									
								
							
							
						
						
									
										10
									
								
								README.md
									
									
									
									
									
								
							@@ -29,10 +29,12 @@ python3 scan.py /path/to/directory [options]
 | 
				
			|||||||
```
 | 
					```
 | 
				
			||||||
 | 
					
 | 
				
			||||||
### Options:
 | 
					### Options:
 | 
				
			||||||
- `-f, --filetype`: Filter by file types (e.g., `.txt`, `.log`).
 | 
					
 | 
				
			||||||
- `-i, --ignore`: Ignore files and folders containing specific strings.
 | 
					To show the help page execute:
 | 
				
			||||||
- `--ignore-hidden`: Omit hidden directories from the scan.
 | 
					
 | 
				
			||||||
- `-v, --verbose`: Enable verbose mode for detailed output.
 | 
					```bash
 | 
				
			||||||
 | 
					python3 scan.py /path/to/directory --help
 | 
				
			||||||
 | 
					```
 | 
				
			||||||
 | 
					
 | 
				
			||||||
## ⚠️ Caution
 | 
					## ⚠️ Caution
 | 
				
			||||||
 | 
					
 | 
				
			||||||
 
 | 
				
			|||||||
							
								
								
									
										26
									
								
								scan.py
									
									
									
									
									
								
							
							
						
						
									
										26
									
								
								scan.py
									
									
									
									
									
								
							@@ -57,19 +57,26 @@ class DirectoryHandler:
 | 
				
			|||||||
            dirs[:] = [d for d in dirs if not d.startswith('.')]
 | 
					            dirs[:] = [d for d in dirs if not d.startswith('.')]
 | 
				
			||||||
        dirs[:] = [d for d in dirs if not any(ig in d for ig in ignore_strings)]
 | 
					        dirs[:] = [d for d in dirs if not any(ig in d for ig in ignore_strings)]
 | 
				
			||||||
 | 
					
 | 
				
			||||||
    @staticmethod
 | 
					    def should_print_file(file_path, file_filters, ignore_strings, ignore_hidden, whitelist_file, whitelist_content):
 | 
				
			||||||
    def should_print_file(file_path, file_filters, ignore_strings, ignore_hidden, include_strings):
 | 
					        """Determine if a file should be printed based on new criteria."""
 | 
				
			||||||
        """Determine if a file should be printed."""
 | 
					 | 
				
			||||||
        if ignore_hidden and os.path.basename(file_path).startswith('.'):
 | 
					        if ignore_hidden and os.path.basename(file_path).startswith('.'):
 | 
				
			||||||
            return False
 | 
					            return False
 | 
				
			||||||
        if file_filters and not any(file_path.endswith(file_type) for file_type in file_filters):
 | 
					        if file_filters and not any(file_path.endswith(file_type) for file_type in file_filters):
 | 
				
			||||||
            return False
 | 
					            return False
 | 
				
			||||||
        if any(ignore_str in file_path for ignore_str in ignore_strings):
 | 
					        if any(ignore_str in file_path for ignore_str in ignore_strings):
 | 
				
			||||||
            return False
 | 
					            return False
 | 
				
			||||||
        if include_strings and not any(include_str in file_path for include_str in include_strings):
 | 
					        if whitelist_file and not any(whitelist_str in file_path for whitelist_str in whitelist_file):
 | 
				
			||||||
            return False
 | 
					            return False
 | 
				
			||||||
 | 
					        if whitelist_content:
 | 
				
			||||||
 | 
					            try:
 | 
				
			||||||
 | 
					                with open(file_path, 'r') as f:
 | 
				
			||||||
 | 
					                    content = f.read()
 | 
				
			||||||
 | 
					                if not any(whitelist_str in content for whitelist_str in whitelist_content):
 | 
				
			||||||
 | 
					                    return False
 | 
				
			||||||
 | 
					            except UnicodeDecodeError:
 | 
				
			||||||
 | 
					                return False
 | 
				
			||||||
        return True
 | 
					        return True
 | 
				
			||||||
 | 
					    
 | 
				
			||||||
    @staticmethod
 | 
					    @staticmethod
 | 
				
			||||||
    def print_file_content(file_path, no_comments, compress):
 | 
					    def print_file_content(file_path, no_comments, compress):
 | 
				
			||||||
        """Print the content of a file."""
 | 
					        """Print the content of a file."""
 | 
				
			||||||
@@ -97,7 +104,7 @@ class DirectoryHandler:
 | 
				
			|||||||
        for root, dirs, files in os.walk(directory):
 | 
					        for root, dirs, files in os.walk(directory):
 | 
				
			||||||
            DirectoryHandler.filter_directories(dirs, kwargs['ignore_strings'], kwargs['ignore_hidden'])
 | 
					            DirectoryHandler.filter_directories(dirs, kwargs['ignore_strings'], kwargs['ignore_hidden'])
 | 
				
			||||||
            for file in files:
 | 
					            for file in files:
 | 
				
			||||||
                if DirectoryHandler.should_print_file(os.path.join(root, file), kwargs['file_filters'], kwargs['ignore_strings'], kwargs['ignore_hidden'], kwargs['strings']):
 | 
					                if DirectoryHandler.should_print_file(os.path.join(root, file), kwargs['file_filters'], kwargs['ignore_strings'], kwargs['ignore_hidden'], kwargs['whitelist_file'], kwargs['whitelist_content']):
 | 
				
			||||||
                    DirectoryHandler.print_file_content(os.path.join(root, file), kwargs['no_comments'], kwargs['compress'])
 | 
					                    DirectoryHandler.print_file_content(os.path.join(root, file), kwargs['no_comments'], kwargs['compress'])
 | 
				
			||||||
                elif kwargs['verbose']:
 | 
					                elif kwargs['verbose']:
 | 
				
			||||||
                    print(f"Skipped file: {file}")
 | 
					                    print(f"Skipped file: {file}")
 | 
				
			||||||
@@ -116,15 +123,16 @@ def main():
 | 
				
			|||||||
    parser.add_argument("--ignore-hidden", action='store_true', help="Ignore hidden directories and files.")
 | 
					    parser.add_argument("--ignore-hidden", action='store_true', help="Ignore hidden directories and files.")
 | 
				
			||||||
    parser.add_argument("-v", "--verbose", action='store_true', help="Enable verbose mode.")
 | 
					    parser.add_argument("-v", "--verbose", action='store_true', help="Enable verbose mode.")
 | 
				
			||||||
    parser.add_argument("--no-comments", action='store_true', help="Remove comments from the displayed content based on file type.")
 | 
					    parser.add_argument("--no-comments", action='store_true', help="Remove comments from the displayed content based on file type.")
 | 
				
			||||||
    parser.add_argument("-s", "--strings", nargs='+', default=[], help="Only display files whose paths contain one of these strings.")
 | 
					 | 
				
			||||||
    parser.add_argument("--compress", action='store_true', help="Compress code (for Python files).")
 | 
					    parser.add_argument("--compress", action='store_true', help="Compress code (for Python files).")
 | 
				
			||||||
 | 
					    parser.add_argument("--whitelist-file", nargs='+', default=[], help="Only display files whose paths contain one of these strings.")
 | 
				
			||||||
 | 
					    parser.add_argument("--whitelist-content", nargs='+', default=[], help="Only display files containing one of these strings in their content.")
 | 
				
			||||||
    args = parser.parse_args()
 | 
					    args = parser.parse_args()
 | 
				
			||||||
    
 | 
					    
 | 
				
			||||||
    for path in args.paths:
 | 
					    for path in args.paths:
 | 
				
			||||||
        if os.path.isdir(path):
 | 
					        if os.path.isdir(path):
 | 
				
			||||||
            DirectoryHandler.handle_directory(path, file_filters=args.filetype, ignore_strings=args.ignore, ignore_hidden=args.ignore_hidden, verbose=args.verbose, no_comments=args.no_comments, compress=args.compress, strings=args.strings)
 | 
					            DirectoryHandler.handle_directory(path, file_filters=args.filetype, ignore_strings=args.ignore, ignore_hidden=args.ignore_hidden, verbose=args.verbose, no_comments=args.no_comments, compress=args.compress, whitelist_file=args.whitelist_file, whitelist_content=args.whitelist_content)
 | 
				
			||||||
        elif os.path.isfile(path):
 | 
					        elif os.path.isfile(path):
 | 
				
			||||||
            if DirectoryHandler.should_print_file(path, file_filters=args.filetype, ignore_strings=args.ignore, ignore_hidden=args.ignore_hidden, include_strings=args.strings):
 | 
					            if DirectoryHandler.should_print_file(path, file_filters=args.filetype, ignore_strings=args.ignore, ignore_hidden=args.ignore_hidden, whitelist_file=args.whitelist_file, whitelist_content=args.whitelist_content):
 | 
				
			||||||
                DirectoryHandler.handle_file(path, file_filters=args.filetype, ignore_strings=args.ignore, ignore_hidden=args.ignore_hidden, no_comments=args.no_comments, compress=args.compress)
 | 
					                DirectoryHandler.handle_file(path, file_filters=args.filetype, ignore_strings=args.ignore, ignore_hidden=args.ignore_hidden, no_comments=args.no_comments, compress=args.compress)
 | 
				
			||||||
        else:
 | 
					        else:
 | 
				
			||||||
            print(f"Error: {path} is neither a valid file nor a directory.")
 | 
					            print(f"Error: {path} is neither a valid file nor a directory.")
 | 
				
			||||||
 
 | 
				
			|||||||
		Reference in New Issue
	
	Block a user