I process documents with Textract and a Lambda Python function reads the fields I need. One of them which varies is the date on the documents; these can be in many formats,
