What architecture I should use for work with many types of documents

I want to use best practice architecture for my python application. The main idea of app is file processing.

I have a multiple types of pdf docs. Each type of pdf doc should be parsed and saved to xls.

I want to build some class (factroy or something like this, I don't know) that takes pdf file, decides how to parse file and save result to xls.

What an architecture I should use?