Kamu

Updated
- Extended flow history:
- start condition update event now holds a snapshot of the start condition
- input dataset trigger now returns a queryable Dataset object instead of simply identifier
- dependent dataset flows should not be launched after Up-To-Date input flow
Fixed
- Zero value handling for `per_page` value in GraphQL
- Flow history: more corrections to show natural time of flow events and keep flow system testable
- Fixed metadata chain scanning regression

Changed
- Using `cargo udeps` to detect unused dependencies during linting
Fixed
- Flow history no longer produces duplicate flow abortion events
- Flow history no longer shows initiation time that is earlier than the next event's time

Changed:
- Flow status `Cancelled` was finally replaced with `Aborted`, unifying cancellation types for simplicity
Fixed
- Flow system now pauses flow configuration, even when cancelling an already completed flow

Added
- Added support of wildcard patterns for `kamu pull` and `kamu push` commands
- Added `{dataset}/tail` and `/query` REST API endpoints
Changed
- Optimization of passes through metadata chain with API using Visitors

Fixed
- Fixed `mapboxgl` dependency issue in Jupyter image

Changed
- Simplified flow statuses within Flow System (no more Queued or Scheduled status)
- Extended flow start conditions with more debug information for UI needs
- Simplified flow cancellation API:
- Cancelling in Waiting/Running states is accepted, and aborts the flow, and it's associated tasks
- Cancelling in Waiting/Running states also automatically pauses flow configuration