Raj rajrao

Download the Power BI MCP Server extension VISX to your machine from: https://analysis-services.gallery.vsassets.io/_apis/public/gallery/publisher/analysis-services/extension/powerbi-modeling-mcp/latest/assetbyname/Microsoft.VisualStudio.Services.VSIXPackage (this is the VISX for: https://marketplace.visualstudio.com/items?itemName=analysis-services.powerbi-modeling-mcp)

Once you have downloaded and extracted the files to a local folder (below example uses "C:\Tools\kiro\"), you can update the mcp.json with the following

{
  "mcpServers": {
    "powerbi-modeling-mcp": {
 "command": "C:\\Tools\\kiro\\extension\\server\\powerbi-modeling-mcp.exe",

name	Calendar Instructions and Examples
description	Guidelines for creating Power BI calendar objects
uriTemplate	resource://calendar_instructions_and_examples

Calendar Column Groups Guide

This guide explains how to define calendar column groups in a Power BI date table so time intelligence works as expected and consistently across models.

Concepts

Mermaid

    F[👆 First Touch<br/>26.7%]:::green 
    L[📝 Lead Creation<br/>26.7%]:::orange 
    O[💼 Opportunity Creation<br/>26.7%]:::purple 
    
    M1[🔄 Middle Touches<br/>10%]:::red
    I1[📢 Impression<br/>10%]:::blue 
    M2[🔄 Middle Touches<br/>10%]:::red
 I2[📢 Impression10%]:::blue

The following query allows you to search for a requestid and get a link to the event (which is obtained by adding @log to the query)

fields @timestamp as Timestamp, @message, @logStream as LogStream, @log
| filter(@requestId = '09d10d9b-e8e0-46fa-ad9e-01a00c87d221')
| sort @timestamp desc
| limit 10

Get stats about a certain kind of message

	import logging
	import os
	from fastmcp.server.middleware import Middleware, MiddlewareContext
	from fastmcp.server.dependencies import get_http_headers
	from starlette.responses import JSONResponse


	_logger = logging.getLogger(__name__)

	class AuthKeyMiddleware(Middleware):

	from requests_toolbelt.utils import dump

	headers = {'Content-Type': 'application/json'}
	response = requests.post(
	url,
	headers=headers,
	json=jdata,
	)
	print(dump.dump_all(response).decode("utf-8"))

	import timeit
	from pprint import pprint

	setup = '''
	from string import Template
	s = "the quick brown fox JUMPED OVER THE"
	t = "LAZY DOG'S BACK 1234567890"
	'''
	iter = 1
	baseline = timeit.timeit("f'{s} {t}'", setup, number=iter)

	def get_backoff_w_jitter(attempt: int, max_backoff: int = 20) -> int:
	"""returns the wait time for a backoff for attempt # attempt
	algorithm: truncated binary exponential backoff with jitter
	src: https://docs.aws.amazon.com/sdkref/latest/guide/feature-retry-behavior.html

	Arguments:
	attempt -- the attempt # for calculating the backoff. a number >= 0

	Returns:
	the backoff amount (usually used as seconds to sleep)

	# based on https://github.com/apache/airflow/blob/9ac742885ffb83c15f7e3dc910b0cf9df073407a/airflow/sensors/base.py#L251C13-L251C13
	import hashlib
	from datetime import datetime, timedelta
	from time import sleep

	poke_interval = 5 # Time in seconds that the job should wait in between each tries
	timeout = 60 # Time, in seconds before the task times out and fails.
	started_at = datetime.now()
	sleep(1)
	exponential_backoff = True

	import os
	cur_working_dir = os.path.join(os.getcwd(), os.path.dirname(__file__))
	print(cur_working_dir)