- Refactor the code to decide how to split `Spectra` for parallel processing:
  it's no longer done automatically by `dataStorage`.
- Add a slot to `Spectra` allowing to set a processing chunk size. Issue #304
- With chunk-wise processing only the data of one chunk is realized to memory in
  each iteration. This also enables to process data in parallel.

- Add and modify functions to enable default chunk-wise processing of peaks
  data (issue #304).
- Split the documentation for chunk-wise (parallel) processing into a separate
  documentation entry.

- Use `processingChunkFactor` instead of `dataStorage` in functions to define
  splitting and processing.
- Remove unnecessary functions.
- Add a vignette on parallel/chunk-wise processing.