This dataset contains one row per programming problem. The split is test.
task_num: LeetCode problem id.
title: Problem title.
difficulty: Numeric difficulty label from the source metadata.
func_name: Name of the function under test.
description: Problem statement text.
solution_code: Full Python source from data/code_<task_num>.py.
tests: List of full Python test-runner source files for the problem, ordered by testcase id.… See the full description on the dataset page:
https://huggingface.co/datasets/wwhhww/llm-path-test-gen.
