CLM - Coding-Academy Lecture Manager

Version: 1.25.0 | License: MIT | Python: 3.12, 3.13, 3.14

CLM is a course content processing system that converts educational materials (Jupyter notebooks, PlantUML diagrams, Draw.io diagrams) into multiple output formats.

# Install from PyPI
pip install coding-academy-lecture-manager

# Or with all optional dependencies (workers, TUI, web dashboard)
pip install "coding-academy-lecture-manager[all]"

For development, clone the repository and install in editable mode:

git clone https://github.com/hoelzl/clm.git
cd clm
pip install -e ".[all]"

# Convert a course
clm build /path/to/course.xml

# Watch for changes and auto-rebuild
clm build /path/to/course.xml --watch

# Show help
clm --help

• Multiple Output Formats: HTML slides, Jupyter notebooks, extracted code
• Multi-Language Notebooks: Python, C++, C#, Java, TypeScript, Markdown
• Diagram Support: PlantUML and Draw.io conversion
• Multiple Output Targets: Separate student/solution/instructor outputs
• Shared-Source Includes: Declare <include source="…" as="…"/> on a <topic> or <section> to splice a canonical Python package (or any file/directory) into multiple topics at build time. clm course sync-includes materializes the same sources on disk so local notebook execution (VS Code, JupyterLab) finds them, with a .clm-include ledger for safe cleanup.
• Output-Write Deduplication: When the same file is written to the same output path by multiple producers, CLM deduplicates the write and surfaces a output_dedup_count / output_conflicts summary so you can spot accidental cross-topic collisions.
• Watch Mode: Auto-rebuild on file changes
• Incremental Builds: Content-based caching
• LLM Summaries: Generate course summaries with clm export summary using any OpenAI-compatible LLM API
• Recording Management: Manage video recording workflows with pluggable backends — local ONNX pipeline, iZotope RX 11 external tool, or Auphonic cloud processing — plus assembly, job tracking, and per-course status (clm recordings)
• MCP Server: Model Context Protocol server for AI-assisted slide authoring (clm mcp) with 16 tools for course navigation, validation, normalization, and bilingual editing
• Slide Authoring Tools: Split-deck authoring sync (clm slides sync — the funnel that keeps both halves of a .de/.en pair consistent), topic resolution (clm course resolve-topic), fuzzy search (clm slides search), spec/slide validation (clm validate), normalization (clm slides normalize), bilingual language view (clm slides language-view), voiceover extraction (clm voiceover extract), LLM-driven voiceover coverage check (clm slides coverage), bilingual ↔ per-language file conversion (clm slides split / clm slides unify), and structured JSON outlines (clm export outline --format json). Lower-level plumbing (clm slides assign-ids, clm slides suggest-sync) stays available by name for scripts and agents.
• Video Narration Harvest: Recover spoken narration from recorded videos into slide decks (clm harvest) — an agent-first report → task → accept loop over the deterministic transcribe/detect/match/align pipeline, with multi-part recordings, caching, and a legacy embedded-LLM one-shot (clm harvest autopilot)
• LLM Polish: Clean up speaker notes with LLM-powered text polishing (clm polish)
• Git Integration: Manage output repos with clm git init/sync/status, including --amend and --force-with-lease for iterative workflows
• Flexible Remote URLs: Configurable git remote URL templates for SSH, custom hosts, etc.

For Users:

• User Guide - Complete usage guide
• Quick Start - Build your first course
• Spec File Reference - Course XML format
• Configuration - Configuration options
• Changelog - Version history

For Developers:

• Contributing Guide - How to contribute
• Developer Guide - Development documentation
• Architecture - System design
• AGENTS.md - AI assistant reference (imported by CLAUDE.md for Claude Code)

# Install pre-commit hooks (recommended)
uv run pre-commit install

# This enables automatic linting (ruff) and type checking (mypy) on every commit

# Run unit tests
pytest

# Run all tests (unit, integration, e2e)
pytest -m ""

# Run with coverage
pytest --cov=src/clm

MIT License - see LICENSE for details.

• Repository: https://github.com/hoelzl/clm/
• Issues: https://github.com/hoelzl/clm/issues