Kamu

Fixed
- Reverted unstable AWS API dependencies

Changed
- Massive crates restructuring around Open Data Fabric code:
- `src/odf` concentrates large number of related crates now, preparing for future separation in different Git repo
- old `opendatafabric` crate becamse `odf-metadata`
- low-level repository implementations became `odf-storage[-...]` crates
- specific storage technologies are gated via features (`lfs`, `s3`, `http`)
- `DatasetFactory`, `Dataset`, `DatasetSummary`, `DatasetLayout`, `BlockRef`,
`MetadataChain` and visiting logic now are residents of `odf-dataset`
- `kamu-data-utils` became `odf-data-utils`
- multiple utility libraries introduced, shared both via ODF and main Kamu code
(`async-utils`, `file-utils`, `s3-utils`, `test-utils`)
- `DatasetRepository` still stays in Kamu core, renamed to `DatasetStorageUnit` to better match it's current purpose
- import statements pointing at ODF code have been cleaned all over the code base (use `odf` meta-crate only outside `src/odf`)

Changed
- Private Datasets:
- OSO: using user actors / dateset resources that come from the database
- Thus, any access check relies on real entities
- GQL, added `Dataset.visibility()` to get the current visibility value
- GQL, added `DatasetMut.setVisibility()` to be able to change the dataset visibility after it has been created
- Deletion of previously created (and unused) ReBAC-properties and reindexing
- OSO: updating the schema to use identifiers instead of names
- OSO: added resource storage for access speed
- E2E: Using the correct account in multi-tenant mode
- And also the possibility of set it up
- `DatasetOwnershipService`: moved to the `kamu-dataset` crate area & implemented via `DatasetEntryServiceImpl`
- GQL, `DatasetMetadata.currentUpstreamDependencies`: indication if datasets not found/not accessed
- GQL, `DatasetMetadata.currentDownstreamDependencies`: exclude datasets that cannot be accessed
- E2E: added the ability to create an account using CLI

Fixed
- Fix crash on resolving dataset by non-existing account
- Minor improvements in event sourcing aggregation

Changed
- Updated to latest `datafusion` and `alloy` dependencies
- Performance improvements with batch loading of event sourcing aggregates

Changed
- Extended database config options with next fields: `maxConnections`, `maxLifeTimeSecs` and `acquireTimeoutSecs`