A custom exception type is usually defined by:

A custom exception type is typically defined by inheriting from the exception class. This method allows you to create a new exception type that can encapsulate specific error conditions relevant to your application or library, providing more context about the error than a standard exception. By defining a custom exception class, you can include additional attributes or methods to better describe the nature of the exception, making it easier to handle and debug.

For instance, when you create a custom exception by subclassing the base exception or one of its descendants, you can tailor the behavior of the exception to suit the needs of your specific application. You might want to add extra data that can help when you catch the exception later on or override certain methods to change how it behaves when printed or logged.

The other options do not reflect the common practice for defining a custom exception. Inheriting from the try block does not make sense as try blocks are used for catching exceptions rather than defining them. Function definitions within the except block would be related to handling exceptions rather than creating custom exception types. Lastly, a class definition within the finally block is also not appropriate because finally blocks are designed for cleanup actions that need to take place after try and except blocks, rather than for defining new exceptions.