wwhhww/llm-path-test-gen

LLM Path Test Generation

This dataset contains one row per programming problem. The split is test.

Fields

• —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.

Python File Organization

solution_code is the complete contents of the corresponding solution file:

data/code_<task_num>.py

Each item in tests is the complete contents of a generated test runner:

run_code_generated/run_<task_num>_<testcase_id>.py

The test runners import Solution from data.code_<task_num>, define run_main(), construct inputs, and call the target method.