Emolysis for Microsoft Fabric.

1. 01

   Choose Fabric

   Sign in to Emolysis, then choose Fabric in the app. From a terminal, the command is emo connect fabric.

2. 02

   Sign in with Microsoft Entra ID

   You sign in with your organization's Microsoft Entra ID account. The credentials are kept in your machine's credential store and never sent to Emolysis.

3. 03

   Pick a workspace

   Emolysis lists the workspaces your account can reach. You pick the workspace and the lakehouse for deploys.

4. 04

   Sample on request

   When you ask for a sample, Emolysis reads a small slice of a table from a lakehouse through Spark into DuckDB on your machine. That query runs on your platform; the tests that follow do not.

$emo connect fabric

1. 01

   Items in the right order

   Fabric items depend on each other, and the API needs them created in order with valid definitions. The profile records this trap, so deploys respect it.

2. 02

   The right language per item

   Spark SQL and PySpark for lakehouse notebooks, T-SQL for warehouse objects. The profile keeps the agent from mixing them.

3. 03

   Through the REST API

   The notebook is created or updated through the Fabric REST API, under your account.

4. 04

   Tagged for Fabric

   Every generated file carries its platform and profile version, so it is never deployed to another platform without an explicit translation step.

5. 05

   Ready for semantic models

   Aggregation / gold tables land as lakehouse tables that semantic models can read.

6. 06

   Your capacity

   Notebooks and pipelines run on your Fabric capacity, on the schedule the pipeline defines, with your account's permissions.

1. 01

   Item definitions and order

   Deploying workspace items through the API needs care with item definitions and the order the items are created in. Items that refer to each other, such as a pipeline that runs a notebook, have to be created in an order that works.

2. 02

   One language per item

   Lakehouse notebooks run Spark SQL and PySpark, warehouses run T-SQL, and real-time data uses KQL. The profile records which one each item uses, so the agent does not mix them.

1. 01Ingest (batch)A notebook or a pipeline copy into lakehouse Delta tables in OneLake.
2. 02Incremental appendA notebook that stores its watermark and runs on a pipeline schedule, or Dataflow Gen2.
3. 03CDC mergeMERGE on lakehouse Delta tables with Spark SQL.
4. 04SCD Type 2MERGE on lakehouse Delta tables from a notebook.
5. 05SnapshotOverwrite or partitioned writes to lakehouse Delta tables.
6. 06BackfillA notebook with parameters, called once per date from a pipeline.
7. 07Late-arrivingA notebook that finds the affected dates and rewrites those partitions.
8. 08Dedupe / conformWindow functions in Spark SQL on lakehouse tables.
9. 09Aggregation / goldLakehouse or warehouse tables that feed semantic models.
10. 10Reverse ETLA pipeline copy activity into a database, or a notebook for APIs.

All ten shapes, with their templates and golden tests, are on the pipeline shapes page.