kedro-datasets: Update __repr__ for PartitionedDataset

[ElenaKhaustova]

Description

The pretty printing implementation relies on a dataset's _describe() method to get the content for printing. For PartitionedDataset _describe() returns dataset type and config which should be formatted similarly to CachedDataset for consistency.

Current representation:

Target representation (see CachedDataset example):

Context

kedro-org/kedro#3987 (comment)

Possible Implementation

Update PartitionedDataset.__repr__ like it's done for CachedDataset to align representations - add dataset key and use parent _pretty_repr method.