- Improved transformation steps (650) (BREAKING) - Split value/formula/function concepts - Renamed a few minor step arguments
4.0.0a11
- Improved layout and data streams concepts (648) (BREAKING) - Renamed `data_stream` to `list_stream` - Renamed `readData` to `readLists` - Renamed `sample` to `fragment` (`sample` now is raw lists) - Implemented loader.buffer - Implemented parser.sample - Added support for function based checks - Added support for function based steps
4.0.0a10
- Reworked Error.tags (BREAKING) - Reworked Check API and split labels/header (BREAKING)
4.0.0a9
- Rebased on `Detector` class (BREAKING) - Migrated all infer_*, sync/patch_schema and detect_encoding parameters to `Detector` - Made `resource.infer` omit empty objects - Added `resource.read_*(size)` argument - Added `resource.labels` property
4.0.0a8
- Improved checks/steps API (621) (BREAKING) - Updated `validate(extra_checks=[...])` to `validate(checks=[{"code": 'code', ...}])`
4.0.0a7
- Updated describe/extract/transform/validate APIs (BREAKING) - Removed `validate_table` (use `validate_resource`) - Removed legacy `Table` and `File` classes - Removed `dataflows` plugin - Replaced `nopool` by `parallel` (not parallel by default) - Renamed `report.tables` to `report.tasks` - Rebased on `report.tasks[].resource` (instead of plain path/scheme/format/etc) - Flatten Pipeline steps signature