mdf-transcript-to-lessons

中文文档

Convert raw transcripts or long course documents into runnable lesson-by-lesson MarkdownFlow scripts with stable lesson boundaries and reusable variable indexing.

Use Cases

• Convert noisy transcript/course docs into per-lesson MarkdownFlow teaching scripts.
• Preserve code blocks and image links while keeping source-span traceability.
• Produce course index + global variable table for downstream multi-lesson orchestration.

Boundaries

• This component outputs lesson artifacts only, not final course publishing assets.
• It does not replace deep instructional review; it prepares runnable lesson prompts.
• It expects text/markdown inputs, not OCR extraction from binary files.

Quickstart (3 minutes)

python3 -m venv .venv
source .venv/bin/activate
pip install -e ".[dev]"
mdf-transcript-to-lessons --input examples/sample_transcript.md --output-dir output
python -m json.tool output/summary.json

Minimal Runnable Example

Input: examples/sample_transcript.md

Generated outputs:

• output/course_index.json
• output/global_variables.json
• output/lessons/lesson-*.md
• output/summary.json

Architecture: Skill Core + Adapters

• Skill Core: src/mdf_transcript_to_lessons/core.py
  • Platform-agnostic transcript-to-lessons transformation.
• CLI Adapter: src/mdf_transcript_to_lessons/cli.py
  • Local executable entrypoint for automation pipelines.
• OpenClaw Adapter: adapters/openclaw
• Claude Adapter: adapters/claude
• Codex Adapter: adapters/codex

Multi-Agent Compatibility

• OpenClaw example: examples/openclaw_call.md
• Claude example: examples/claude_call.md
• Codex example: examples/codex_call.md

AI-Shifu Ecosystem

This repository is one reusable course-production component in AI-Shifu ecosystem.

• Website: https://ai-shifu.com

Development

pip install -e ".[dev]"
ruff check .
pytest -q