Kamu

Added
- Smart Transfer Protocol now supports pushing into local filesystem workspace repository.
It can be activated via:
`kamu push <dataset_name> --to odf+http://<server-address>:<port>/<dataset_name>`.
I.e., run API server in one Kamu workspace directory, where you intend to push `my-dataset`to:
`kamu system api-server --http-port=35433`
and push from the Kamu workspace directory, where `my-dataset` is stored:
`kamu push my-dataset --to odf+http://localhost::35433/my-dataset`
Changed
- Upgraded to latest `datafusion` and `arrow`
- Updated to multi-tenant references and aliases in accordance with ODF spec
- New datasets no longer use staging area before `DatasetBuilder` finishes them

Added
- Smart Transfer Protocol: implemented Push flow for S3-based dataset repository only.
Does not work with the local workspace yet.

Fixed
- Updated IPFS gateway status code handling after upstream fixes to correctly report "not found" status (108)

Changed
- Revised workspace dependencies management:
- Sharing single definition of common dependencies between modules
- Using `cargo-deny` utility for dependencies linting
- Migrated to official S3 SDK (got rid of unmaintained `Rusoto` package)
- Moved developer's guide to this repository

Added
- Smart Transfer Protocol: implemented Pull flow based on web-sockets. It can be activated via:
`kamu pull odf+http://<server-address>:<port>/<dataset_name>`.
I.e., run API server in one Kamu workspace directory, where a `my-dataset` dataset is present:
`kamu system api-server --http-port=35433`
and pull from another Kamu workspace directory:
`kamu pull odf+http://localhost::35433/my-dataset`
- S3-based dataset repository implementation
Changed
- Improved status codes and error handling in API server's routes for Simple Transfer Protocol

Changed
- Updated to new `rustc`
- Updated to latest `datafusion` and `arrow`
- Improved release procedure