gokart

What problem does it solve?

It reduces bugs and maintenance costs in gokart machine-learning pipeline code by enforcing type-safe task design, Pandera schema validation, and consistent testing patterns that make pipelines reliable and reviewable.

Core Features & Use Cases

• Type safety guidance: Patterns for using TaskOnKart[T], TaskInstanceParameter type annotations, and instance-based loading to retain static typing.
• Pandera integration: How to declare DataFrame[Schema], validate outputs, and design schema inheritance for robust DataFrame contracts.
• Testing & review practices: Test patterns with test_run/build, minimal mocking rules, and a detailed review checklist for code review workflows.
• Use case: Reviewing or implementing a gokart pipeline that merges multiple Pandera-typed datasets, validates outputs, and is covered by unit tests.

Quick Start

Review a gokart task to ensure generic TaskOnKart[T] annotations are present, TaskInstanceParameter parameters are type-annotated, self.load is called with task instances (not string keys), and outputs use Pandera DataFrame[Schema] for validation.