Set up MetricFlow

Getting started

First, if you want to follow along, we'll need to clone the example project. You will need access to a Snowflake, BigQuery, Databricks, or Postgres warehouse for this, for the time being. The project is our classic Jaffle Shop, a simulated chain restaurant serving jaffles and tasty beverages.

git clone git@github.com:dbt-labs/jaffle-sl-template.git
cd path/to/project

Next, before you start writing code, you need to install MetricFlow as an extension of a dbt adapter from PyPI (dbt Core users only). The MetricFlow is compatible with Python versions 3.8 through 3.11.

We'll use pip to install MetricFlow and our dbt adapter:

# activate a virtual environment for your project,
# if you don't have a name you like to use we suggest .venv
python -m venv [virtual environment name]
source [virtual environment name]/bin/activate
# install dbt and MetricFlow
pip install "dbt-metricflow[adapter name]"
# e.g. pip install "dbt-metricflow[snowflake]"

Lastly, to get to the pre-Semantic Layer starting state, checkout the start-here branch.

git checkout start-here

For more information, refer to the MetricFlow commands or a quickstart to get more familiar with setting up a dbt project.

Basic commands

• 💻 This package will install both dbt and mf as CLIs in our virtual environment. All the regular dbt commands like run, build, and test are available.
• 🔍 A less common one that will come in handy with the Semantic Layer is dbt parse. This will parse your project and generate a semantic manifest, a representation of meaningful connections described by your project. This file gives MetricFlow a state of the world from which to generate queries.
• 🧰 In addition to dbt, you'll have access to mf commands like query and validate-configs, which operate based on that semantic manifest. We'll dig more into all of these as we go along.
• 🛠️ Lets start off by running a dbt build to get the starting state of our project built.