Add Cache via Redis

Add a mechanism to cache data which can be identified by an identifier "column"/key in a redis db. Additionally, it extends the HERODataarchiver to act as sink for camera/atomiq etc.

For the cache, we assume that the data has a loose table-like structure. I.e. each payload carries an identifier stored in a field with a known name (default field name identifier). Each unique identifier marks one "row" of the data. To handle several tables, we additionally include a grouping_key which can be for example the run id of an atomiq run.

Resulting from the table-like structure, we can perform operations on rows (via the identifiers) or columns (via column names).