Add support for remote paths for raw data and cellpy data access

[msudars]

• Loading raw data from remote locations
• Searching/listing remote raw-data directories
• Reading and possibly writing cellpy files stored remotely
• handling of authentication and credentials
• when direct remote access is not possible: Predictable local caching / temporary copy behavior

Questions to resolve

• Which remote schemes do we want to support initially?
  • ssh://
  • sftp://
  • scp://
  • others?
• What happens if the remote connection isn't accessible to the user? Additional permissions necessary?
• Should support be limited to raw data first?
• How should credentials be provided?
  - environment variables
  - SSH config / agent
  - key file path
  - password fallback
• What should the expected behavior be for:
  - exists()
  - is_file()
  - glob() / rglob()
  - copying to local temp storage
  - saving back to remote

Acceptance criteria

• A user can configure a remote raw-data directory and load supported raw files from it.
• Remote raw-file discovery works in a documented and tested way.
• Credential handling is documented and tested.
• Behavior for remote cellpy file loading/saving is explicitly supported or explicitly rejected with a clear error.
• The supported remote-path workflows are covered by tests and documented in the user docs.

Notes

This issue is about remote path support as a feature. Refactoring the internal path abstraction should be treated as a separate issue unless it is
strictly necessary to complete this work.

[jepegit]

This is linked to #371