Refactor RedisItemWriter and RedisItemReader logic

Hello,

I believe there might be a better way to implement the RedisItemReader and RedisItemWriter. Currently, the RedisItemReader stores the Redis data's value in the chunk. The RedisItemWriter then uses an ItemKeyMapper to derive the key from this value and executes the writeKeyValue function with the key and value.

However, I find it quite cumbersome to implement an ItemKeyMapper that derives the key from the value. In a key-value structure, retrieving the key based on the value seems to underutilize the key-value paradigm.

When I was setting up a batch job to delete specific Redis data, I found it inefficient to implement an ItemKeyMapper that derives the key from the value, so I customized the code accordingly. Instead, I implemented the ItemKeyMapper to return the value when given a key.

Therefore, I believe it would be more efficient for the RedisItemReader to store the Redis data's keys in the chunk. If my suggestion is not accepted, I would be very interested in understanding the reasoning behind that decision.

Thank you for considering my suggestion.

Start by locating RedisItemReader, RedisItemWriter, and ItemKeyMapper, then inspect how keys and values move through the chunk and into writeKeyValue. Done means the reader and writer use a consistent key/value contract that supports the described Redis deletion use case.