Is it possible to capture specific parts of a PDF text with AWS Textract