In some situations (e.g. if your company uses some sort of proxy service to access the database rather than providing users with the access credentials directly), you may want to manage your own ETL job. In this case, you can write your own build_script.py using databuilder.extractor.Extractor objects (following amundsen's design patterns) and manage your own ETL process. See build_script.py and metaframe/__init__.py for an example of how to write such a script.
Once this is written, you can bake this script into metaframe by navigating to the repo base directory, deleting the /dist directory, and running the following command:
make && make install
We don't explicitly add an alias for the mf binary, so you'll want to either add ~/.metaframe/bin/ to your PATH, or add the following alias to your .bash_profile or .zshrc file.
alias mf=~/.metaframe/bin/mf
This will allow mf etl to run your custom script.