Making Fabric easier to use
Write the Python, SQL and native Power BI your team already knows. Weaver keeps Fabric under control.
Free and open source. Installs with pip. No new language to learn.
What a Weaver project looks like
Ordinary Python, SQL and native Power BI files, organised by Fabric item. Start with one item; connect more as you need them.
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/
End-to-end DAG orchestration across Fabric workspace
Keep your work. Lose the repeated engineering.
Lakehouse
Python and Spark SQL are both first-class.
Use them together across files, Tables and transformations. Weaver keeps everything connected and running in the right order, without a tangled web of notebooks and pipelines.
Explore Lakehouse →Warehouse
Write the T-SQL your team already knows.
No macros or handcrafted loading procedures. Weaver takes care of staging, change processing, dependencies and the operational plumbing needed for serious data engineering.
Explore Warehouse →Semantic Models
Keep native TMDL.
Apply consistent modelling conventions across your organisation's models. Test DAX calculations against independent expectations, so you can have confidence in the numbers.
Explore Semantic Models →Reports
One Power BI Report project. Different environments and models.
Deploy the same source to different workspaces, with Weaver resolving and verifying the right semantic-model binding for each destination.
Explore Reports →Start with what you already have
Have a CSV, a T-SQL query or a Power BI project?
Start there. You don't need to design an entire data estate before getting something useful running.
Add more as you need it.
Know what ran. Know what needs attention.
When something fails, you shouldn't have to piece together what happened from pipeline histories, notebooks and log files.
Weaver keeps track of dependencies, data changes, test results and failures together. You can see what's current, what's stale and where to pick up when something goes wrong.
Multiple environments, without the variable tangle
Keep one set of Python, SQL and Power BI files across development and production.
Weaver handles the different Fabric item names and model bindings without burying your code in environment-specific variables.
And with Mirror, development can use existing data without copying everything first.
Useful at three objects. Engineered for three thousand.
Repeatable deployment and testing are useful even for a couple of files and a Table.
As your estate grows, Weaver takes care of increasingly difficult engineering: dependencies, incremental processing, scheduling, concurrency and recovery.
The same approach works whether you're managing three objects or three thousand.
In a synthetic 3,000-object test estate on F64, the first load completed in about an hour.