We had difficulties deciding between two main challenges for this chapter. We could add additional plugins or refine what currently exists. In actual development, your time would be spent balancing these two objectives as the framework continues to grow. For this chapter, we propose a recursive-based challenge.
Remember that, while explaining the post Office 2007 format of documents, we determined that attached media is stored in the media subdirectory of the document. In its current incarnation, when an Office document is encountered, that media subdirectory, which might have copies of files containing embedded metadata themselves, isn't processed. The challenge here is to add the newly discovered files to the current file listing.
We might do that by returning a list of newly discovered files back to metadata_parser.py. Another route might be to check the file extensions in the office_parser.py script and pass them immediately onto the appropriate plugins. The latter method would be easier to implement but not ideal as it removes some of the control from the metadata_parser.py script. Ultimately, it's up to the developer to determine the most efficient and logical method of completing this challenge.
Beyond this, some other efficiency achievements can be made. For example, we don't need to return the headers for the plugin each and every time the plugin is called. Since the headers will always be the same, we only need to have them created/returned once. Alternatively, this framework is limited by the types of writers it supports. Consider adding a writer for Excel spreadsheets to create more useful reports.