Upload-based metadata editing and refactored upload and entry metadata.

This combines #580 (closed) with a new edit API.

There are lot of duplicated metadata in our mongo collections for processing.Upload and processing.Calc. Originally, the search index and archive is only updated from Calc.metadata and therefore, many upload properties are copied into Calc.metadata. This should be changed.

• remove duplicated and redundant metadata from processing.Calc #580 (closed)
• align metadata style of Upload and Calc. E.g. use metadata dict in both classes. Also #580 (closed)
• add derived property or function (e.g. all_metadata) to processing.Calc that produces a merged dict of upload and entry metadata
• create EntryMetadata (and search index indirectly) from these combined dicts in processing.data.py
• replace all specific search index update operations with one function that is based on these merge metadata dicts
• improve POST uploads/{id}/metadata. It should run a process that performs all necessary Upload/Calc and search index updates. This will later replace the existing edit API. This also needs to validate the metadata, have a respective pydantic model, etc.