Source: 📖 Effective Python item 20

Returning None can be error prone

If a function can return a value of None, this can lead to errors when the function return is evaluated as part of a logical expression. If the caller knows that None is a possible output of the function, this may be used to test for false-equivalent values in an if statement, which would include not only None but also return values of 0, False, empty strings/lists/tuples etc.

This can lead to false-negatives occurring, where an if statement that the caller used intending to check for a return value of None does not discriminate between None and [], for example.

Instead of returning None, raise an exception and include documentation to explain specifically what will trigger the exception to be raised. This will allow the caller to handle the exception outside of the function in whatever way is needed, thus reducing possible errors in the code.