How can we configure the Ruff linter to ignore environment-provided global variables? I have a python codebase which contains notebooks, that run in a databricks environment. The environment provides globals such as spark. A perfectly valid example in that context would be: # Databricks notebook source df = spark.sql("SELECT 'hello world'") df.show() ruff check raises a F821 undefined-name error on this minimal example. We could circumvent the problem by Explicitly loading the specific globals. This adds unnecessary complexity since the code will only ever run where the globals are present. Clutter the code with # noqa: F821 Are there other options? Continue reading...
