The community member is asking about the requirements for a custom metadata extractor. A comment suggests that the requirements are to implement or extend the base class found at a specific GitHub link. There is no explicitly marked answer in the provided information.
