Remove need for direct auth with S3

[phwissmann]

Currently, the ingestor is using credentials for the upload. In order to avoid having credentials stored, the archiver backend server should provide presigned urls to upload data

• Change s3 upload to use presigned urls form archiver backend
• create separate s3 upload module
• verify error handling
• add checksum verification
• use openapi based client
• add retries
• expose parameters (chunksize, concurrent files, pool size) and find sensible defaults
• add expiration date to presigned urls
• verify full files by incremental hashing