Multipart upload of large files

[mesemus]

Is your feature request related to a problem? Please describe.

A user wants to upload a large file, primarily using a commandline client or API to the repository.
The current process is using the file upload API, which:

• is slow for large files, as the file is uploaded in a single thread/request
• it is prone to network errors and whenever this occurs, the user has to re-upload the whole file

Describe the solution you'd like

Invenio provides an API/client to enable uploading chunks of large data. At the beginning of the upload,
the client/user will determine the number of chunks. Invenio will provide a URL for each of those chunks
which client/user will use to perform a PUT request to upload the content of the chunk.

Describe alternatives you've considered

The file might already be present on the target storage - in this case, a linking should be used instead
of re-transfer.

Additional context

See the S3 multipart upload documentation at https://docs.aws.amazon.com/AmazonS3/latest/userguide/mpuoverview.html
for details how it is done in S3 protocol.