日本語版

Spec to Cases

This pipeline ingests specification Markdown, combines LLM drafting with rule-based formatting, and outputs validated test cases in JSON. JSON Schema keeps the structure compatible with existing test management and automation frameworks.

Highlights

• Post-process the LLM draft to guarantee JSON that complies with schema.json.
• Preserve key testing attributes such as steps, expected results, and priority through type-preserving transforms.
• Batch-convert Markdown into JSON with CLI scripts.

Key Artifacts

• spec.sample.md — Sample specification input.
• cases.sample.json — Generated test cases.
• schema.json — Validation schema for the output JSON.
• scripts/convert.py — CLI entry point for the conversion.

How to Reproduce

1. In projects/01-spec2cases-md2json/, install required Python dependencies such as jsonschema.
2. Run scripts/convert.py --spec spec.sample.md --output cases.sample.json to regenerate the sample output.
3. The CLI validates the generated content; mismatches against the schema will raise an error.

Next Steps

• Customize LLM prompts to switch templates by domain (API / UI / non-functional).
• Push the case JSON to existing test management tools (e.g., Xray, TestRail) via their APIs.
• Track adoption logs in the weekly summary for future improvements.