Weaver¶
Making Fabric easier to use, and safer to scale.
Weaver projects use ordinary Python, SQL and native model definitions. Weaver handles the installation and load work around those files, and records the results so you can see what ran and what needs attention. The guides begin with one Fabric item, then explain how several items work together.
A Weaver project at a glance¶
A Weaver project is ordinary Python, SQL, TMDL and configuration files organised by Fabric item. The Parcel delivery example below combines a Lakehouse, Warehouse, semantic model and Report.
Parcel example project
parcel/
├── workspace-config.yml
├── workflow.yml
├── Environment/Weaver.Environment/
├── Lakehouse/Landing/
│ ├── Files/Parcel__Events.py
│ ├── Tables/Parcel__Event.py
│ └── Tables/Parcel.CurrentStatus.sql
├── Warehouse/Operations/
│ ├── Parcel.StatusEvent.sql
│ ├── shortcuts.yml
│ ├── Parcel.LakehouseCurrentStatus.sql
│ └── tests/
└── PowerBI/Parcel/
├── Parcel.SemanticModel/
└── Parcel.Report/
Dependencies connect the work; Weaver handles building, loading, testing and tracking its state. Projects and estates · Connect items and dependencies.
- Lakehouse: files and Python or Spark SQL transformations, with managed Tables and dependency ordering.
- Warehouse: T-SQL SELECTs with the Table structure and load machinery generated around them.
- Semantic models: native TMDL deployment, modelling conventions and DAX validation.
- Reports: native Report deployment and model bindings for each workspace.
When those pieces need to work together, Coordinate your estate covers dependencies, environments, Health and the development loop. Two items are enough to benefit.
New to Weaver? Start here.