Semantic Link for Microsoft Fabric (SemPy)

0.14.0 · active · verified Sat Apr 11

Semantic link connects Microsoft Fabric's Power BI semantic models with data science tools. SemPy is the Python library that is part of semantic link, providing functionalities to analyze and manipulate data from Power BI models directly within notebooks. It integrates deeply with Fabric DataFrames and allows for tasks like listing Power BI items, reading tables, and calculating measures. The current version is 0.14.0, and it's actively developed by Microsoft.

Warnings

gotcha SemPy is primarily designed for and best utilized within Microsoft Fabric notebooks. Running its Fabric-specific functionalities (e.g., `sempy.fabric` functions) outside of this environment will typically fail due to authentication or environment dependencies.
Fix: Run your SemPy code in a Microsoft Fabric notebook for full functionality. Ensure your environment has the necessary authentication configured if attempting to use outside Fabric.
gotcha The library primarily interacts with Power BI semantic models and their underlying data. It's not a general-purpose data manipulation library for arbitrary data sources, although it can work with pandas DataFrames as an entry point.
Fix: Understand that SemPy's strength lies in its integration with Power BI semantic models. For general data tasks, consider using pandas directly or other libraries.
breaking As a pre-1.0 library (version 0.x.x), SemPy's API might experience breaking changes between minor versions. Always check the official documentation or release notes when upgrading.
Fix: Pin your `semantic-link-sempy` version in your `requirements.txt` (`semantic-link-sempy==0.14.0`) and thoroughly test your code after any upgrades.
gotcha SemPy requires Python 3.10 or newer. Using older Python versions will result in installation or runtime errors.
Fix: Ensure your Python environment is version 3.10 or higher. Upgrade your Python interpreter if necessary.

Install

pip install semantic-link-sempy Install latest version

Imports

FabricDataFrame
```
from sempy.fabric import FabricDataFrame
```
FabricDataFrame is part of the 'fabric' submodule, not directly under 'sempy'.
list_workspaces
```
from sempy.fabric import list_workspaces
```
Functions interacting with Fabric entities are typically found in the 'sempy.fabric' module.
read_table
```
from sempy.fabric import read_table
```
Data reading functions for Fabric are in 'sempy.fabric'.

Quickstart

This quickstart demonstrates how to create a `FabricDataFrame` from a standard pandas DataFrame. It then attempts to use `sempy.fabric.list_workspaces()` to show interaction with Microsoft Fabric. Note that most `sempy.fabric` functions require execution within a Microsoft Fabric notebook environment with an authenticated session to function correctly.

import pandas as pd
from sempy.fabric import FabricDataFrame, list_workspaces

# Create a sample pandas DataFrame
df_pd = pd.DataFrame({
    'ID': [1, 2, 3],
    'Value': [100, 200, 150]
})

# Convert a pandas DataFrame to a FabricDataFrame
df_sempy = FabricDataFrame(df_pd)
print("--- FabricDataFrame created ---")
print(df_sempy.head())
print(f"Type: {type(df_sempy)}\n")

# Attempt to list Fabric workspaces
# This function (and most of sempy.fabric) requires an active Microsoft Fabric notebook environment
# and an authenticated session to execute successfully.
print("--- Attempting to list Fabric workspaces ---")
try:
    workspaces = list_workspaces()
    print(f"Successfully listed {len(workspaces)} workspaces.")
    if workspaces:
        print(f"First workspace: {workspaces[0]['display_name']}")
except Exception as e:
    print(f"Failed to list workspaces. This is expected if not running in a Fabric notebook or without proper authentication.\nError: {e}")

view raw JSON →