AS CODE

Your whole Fabric estate, in one repo.

Workspaces, lakehouses, connections, tabular models, notebooks, reports — every part of your platform lives as YAML in a single repository. Git history is your audit trail. Branches are your environments. Your platform is finally text you can read, review, and diff. Machine-readable means AI-readable.

Nested folders: Dataplatform contains DP, which contains Configuration, Connections, Models, Notebooks, Objects, and Reports.

SINGLE SOURCE

Everything is generated from YAML. One place for all your changes.

Add a column to model.yaml — the gold table and tabular model regenerate themselves. Update an objects/*.yaml — bronze and silver follow. One change, propagated through every layer. No manual rebuilds. No schema drift between environments.

ARCHITECTURE

No updates, just inserts. Primary keys validated and enforced.

Every record is an insert with a timestamp. Deletes are negative-timestamp copies. The full history is always there. Primary keys are validated at load time — duplicate or missing keys never reach silver. A DBO view gives you the latest state when you need actuals.

Query results showing product records with productid, productname, category, and a SYSTEMSTATETIMESTAMP column — every row is a timestamped insert.

WHEEL PACKAGE

One Python wheel, driven by your YAML, loads your data.

Every notebook calls the same versioned Python wheel. The wheel reads your YAML to know what to load, how to key it, and where to write it. No hand-rolled ingestion scripts. No copy-paste notebooks. Upgrade once — every notebook follows.

D_Product.ipynb
# One pip install. Always latest version.
!pip install https://publicstor7305.blob.core.windows.net/\
    openwheelpackages/whl/easyfabric-0.0-py3-none-any.whl

from easyfabric import load_data_silver, ConfigManager

# YAML tells the wheel everything it needs to know.
config = {
    'dict_table_config': {
        'path_object_file': 'Files/Model/DM/model.yaml',
    },
    'dict_load_config': {
        'destination_schema': 'dbo',
        'destination_table': 'd_product',
    },
}

load_data_silver(spark, df_result, config)

AI WORKFLOW

Write a story. AI ships it.

Skills for the most common changes ship with the platform — just ask. A real loop: write a story with a skill → it lands in Azure DevOps Boards → Claude breaks it into tasks → Claude executes them → a pull request shows up for review. You stay in control with the merge button.

REPORTS

From YAML to deployed Power BI report. A to Z.

The same pipeline that provisions workspaces and loads data also deploys your Power BI reports. No separate publishing step. No manual import. Acceptance and production roll out from a single commit. End-to-end means end-to-end.

Your Fabric platform, as code. See it on yours.

Book a 30-minute walkthrough with the team — bring your environment, leave with a plan.

Book a demo Read the docs