Skip to main content

A curriculum engine that turns a YAML curriculum definition into a deployable SvelteKit learning application.

Project description

learningfoundry

License Python CI codecov

A curriculum engine that turns a YAML curriculum definition into a deployable SvelteKit learning application — with interactive assessments, executable notebooks, and data visualizations — in a single pipeline.


Table of Contents


Overview

learningfoundry takes a single curriculum.yml file and generates a fully self-contained SvelteKit learning application. The generated app supports:

  • Text — Markdown content rendered in the browser
  • Video — YouTube embeds
  • Quiz — Interactive assessments via quizazz (optional)
  • Exercise — Executable notebooks via nbfoundry (stub provided)
  • Visualization — D3-based charts via d3foundry (stub provided)

Learner progress is persisted locally in SQLite (via sql.js) — no backend required.


Installation

pip install learningfoundry

With optional quizazz support:

pip install "learningfoundry[quizazz]"

Requirements:

  • Python 3.12+
  • pnpm (for preview command and generated app development)
  • Node.js 18+ (for the generated SvelteKit app)

Quick Start

  1. Create a curriculum file (see Curriculum YAML Format):

    cat > curriculum.yml << 'EOF'
    version: "1.0.0"
    curriculum:
      title: "My Course"
      description: "A short description."
      modules:
        - id: mod-01
          title: "Module One"
          lessons:
            - id: lesson-01
              title: "Getting Started"
              content_blocks:
                - type: text
                  ref: content/lesson-01.md
                - type: video
                  url: "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
    EOF
    
  2. Validate the curriculum:

    learningfoundry validate
    # OK — curriculum is valid.
    
  3. Build and preview locally:

    learningfoundry preview
    # Preview server started at http://localhost:5173
    

    learningfoundry preview is the canonical "see your work" command — it builds the SvelteKit project, installs Node dependencies on first run (and again whenever they change), and starts a Vite dev server. On subsequent runs it skips the install step automatically.

    learningfoundry build alone is also available if you want to generate the SvelteKit project without serving it (e.g. to inspect output, deploy a static export via cd dist && pnpm build, or wire into your own toolchain).


CLI Reference

learningfoundry build

Parse → resolve → generate a SvelteKit project.

Usage: learningfoundry build [OPTIONS]

Options:
  -c, --config PATH       Path to the curriculum YAML file.  [default: curriculum.yml]
  --log-level LEVEL       Logging verbosity.  [default: INFO]
                          Choices: DEBUG, INFO, WARNING, ERROR
  -o, --output PATH       Output directory for the generated SvelteKit project.
                          [default: dist]
  --base-dir PATH         Base directory for content refs.
                          (default: curriculum file's parent directory)
  --help                  Show this message and exit.

Exit codes:

Code Meaning
0 Success
1 Curriculum validation error
2 Content resolution error (missing file, bad URL, etc.)
3 SvelteKit generation error
4 Configuration file error

learningfoundry validate

Validate a curriculum YAML without generating any output.

Usage: learningfoundry validate [OPTIONS]

Options:
  -c, --config PATH       Path to the curriculum YAML file.  [default: curriculum.yml]
  --log-level LEVEL       Logging verbosity.  [default: INFO]
  --base-dir PATH         Base directory for resolving content refs.
  --help                  Show this message and exit.

Prints OK — curriculum is valid. on success, or a list of errors and exits with code 1.


learningfoundry preview

Build then launch a local Vite dev server.

Usage: learningfoundry preview [OPTIONS]

Options:
  -c, --config PATH       Path to the curriculum YAML file.  [default: curriculum.yml]
  --log-level LEVEL       Logging verbosity.  [default: INFO]
  -o, --output PATH       Output directory for the generated SvelteKit project.
                          [default: dist]
  --base-dir PATH         Base directory for content refs.
  --port INTEGER          Port for the local dev server.  [default: 5173]
  --help                  Show this message and exit.

Runs learningfoundry build, then pnpm install (skipped when every declared dependency is already present in node_modules/), then pnpm run dev in the generated project directory. Requires pnpm on PATH.

This serves the SvelteKit project from source via Vite's dev server; it does not serve the static pnpm build output in dist/build/. For static deploys, use cd dist && pnpm build and host the resulting dist/build/ directory on any static host.


Curriculum YAML Format

version: "1.0.0"

curriculum:
  title: "Course Title"           # required
  description: "Course overview." # optional

  modules:
    - id: mod-01                  # required, kebab-case
      title: "Module One"         # required
      description: "..."          # optional

      # Optional pre/post assessments (requires quizazz-builder)
      pre_assessment:
        source: quizazz
        ref: assessments/mod-01-pre.yml

      post_assessment:
        source: quizazz
        ref: assessments/mod-01-post.yml

      lessons:
        - id: lesson-01           # required, kebab-case; unique within module
          title: "Lesson One"     # required

          content_blocks:

            # Text block — Markdown file
            - type: text
              ref: content/mod-01/lesson-01.md

            # Video block — `provider` selects the player (default: youtube)
            - type: video
              url: "https://www.youtube.com/watch?v=XXXXXXXXXXX"
              # provider: youtube          # optional today; only youtube is implemented
              # extensions: {}            # optional; player-specific payload (see "Video blocks")

            # Quiz block — requires learningfoundry[quizazz]
            - type: quiz
              source: quizazz
              ref: assessments/mod-01-quiz.yml

            # Exercise block — requires nbfoundry (stub included)
            - type: exercise
              source: nbfoundry
              ref: exercises/mod-01-exercise.yml

            # Visualization block — requires d3foundry (stub included)
            - type: visualization
              source: d3foundry
              ref: visualizations/mod-01-vis.yml

Rules:

  • Module and lesson id values must be unique within their scope, and match the pattern [a-z0-9][a-z0-9-]*.
  • Every curriculum must have at least one module; every module at least one lesson.
  • All ref paths are resolved relative to --base-dir (default: directory containing the curriculum YAML).
  • Only YouTube URLs are accepted for video blocks when provider is youtube (the default): youtube.com/watch?v= or youtu.be/.

Video blocks

Each video content block carries:

  • url — Watch URL for the provider (validated for YouTube when provider: youtube).
  • provider — Which player to use. Omitted in YAML means youtube. New providers (e.g. Vimeo) will add new literal values here together with resolver + frontend support.
  • extensions — Optional mapping of player-specific data. There is no cross-player generic schema: keys and shapes are defined per provider. Examples you might add later for YouTube: chapters (timestamp + title list), transcript_ref (path to WebVTT or plain text), autoplay. The build passes extensions through to curriculum.json unchanged; the Svelte app can grow per-provider components that read content.extensions.

Older generated apps only had url in each video block’s content; the template still treats missing provider as youtube.


Lesson titles and markdown headings

Each lesson page renders two title strings, from two different sources:

  1. The lesson title from curriculum.yml (the title: on a lesson). Used by the sidebar, the breadcrumb, the browser tab, and the page's outer <h1>.
  2. The leading heading in the lesson's markdown file (the # Heading at the top, if any). Rendered inside the lesson body.

If both strings are identical, the page renders the same title twice and looks broken. The fix is purely an authoring convention — there is no rendering bug to chase.

Convention

  • Keep the YAML title: short and navigation-shaped. Either a number ("3"), a label ("Lesson 3"), or label-plus-abbreviation ("Lesson 3: Cultural Diffusion").
  • Make the markdown # Heading the descriptive long-form title that complements the YAML title — never echoes it. Imagine reading them together as "<yaml title>: <markdown H1>"; that sentence should flow naturally and contain no repeated words.
  • If the lesson genuinely has nothing extra to add in a heading, omit the markdown # Heading entirely and start the lesson with body prose. The page already has the YAML title rendered as its <h1>.

Examples

Good — complementary, reads as one sentence:

# curriculum.yml
- id: lesson-03
  title: "Lesson 3"
<!-- content/mod-01/lesson-03.md -->
# The Diffusion of Cultural Artifacts

Most cultural products fail. A reasonable estimate places the fraction…

Renders as:

Lesson 3

The Diffusion of Cultural Artifacts

Most cultural products fail. …

Good — slightly more YAML detail, still no echo:

- id: lesson-03
  title: "Lesson 3: Cultural Diffusion"
# Why Most Pop Releases Disappear

Bad — duplicative; both titles render the same string:

- id: lesson-03
  title: "The Diffusion of Cultural Artifacts"
# The Diffusion of Cultural Artifacts

Also fine — no markdown heading at all:

- id: lesson-03
  title: "Lesson 3: Cultural Diffusion"
Most cultural products fail. A reasonable estimate places the fraction…

Images and assets

Lesson markdown can embed images directly. Place the image file alongside the markdown that uses it and reference it with a relative path:

content/
└── mod-01/
    ├── lesson-01.md
    ├── diagram.png
    └── figures/
        └── architecture.svg
# Lesson One

![Architecture diagram](figures/architecture.svg "Hover title")

Here is a smaller inline diagram:

<img src="diagram.png" alt="Diagram" />

How it works:

  • Relative URLs (diagram.png, figures/architecture.svg) are resolved against the markdown file's own directory. learningfoundry build copies each unique image into dist/static/content/<sha256[:12]>/<basename> and rewrites the markdown URL to the absolute path /content/<sha256[:12]>/<basename> so it resolves at every nested route in the generated app.
  • Both the markdown form (![alt](path), ![alt](path "title")) and the HTML form (<img src="path">) are recognised.
  • Absolute URLs (https://, http://, protocol-relative //..., root-absolute /...) and data: URIs pass through unchanged — useful for CDN-hosted assets you don't want copied into the build.
  • Image references inside fenced code blocks (``` or ~~~) are left as literal text, so code samples that demonstrate image syntax aren't silently rewritten.
  • The same image referenced from N lessons is copied exactly once (deduped by content hash).
  • A missing image fails the build with the lesson location and the expected on-disk path in the error message.

For production deployment to a CDN, just run cd dist && pnpm build — the static/content/ tree gets bundled into the static export under build/content/, so deploying build/ to any static host (Cloudflare Pages, Netlify, S3+CloudFront, …) serves the images at the same URLs the markdown references.


Content locking

Control access to modules and lessons with a three-level configuration hierarchy (most local wins):

  1. Per-module locked — explicit true/false override; trumps everything.
  2. Curriculum locking.sequential — when true, module N+1 requires module N complete.
  3. Global config locking.sequential — project-wide default (see Configuration File below).
curriculum:
  locking:
    sequential: true            # modules must be completed in order
    lesson_sequential: false    # lessons within a module are free-order

  modules:
    - id: mod-01
      locked: false             # always accessible regardless of sequential
      lessons:
        - id: lesson-01
          unlock_module_on_complete: true   # completing this unlocks siblings + next module
          content_blocks:
            - type: quiz
              source: quizazz
              ref: assessments/quiz.yml
              pass_threshold: 0.7           # 70% required to count as passed

unlock_module_on_complete is useful for "gateway" lessons — a single assessment that, once passed, opens the rest of the module and the next one.


Configuration File

An optional config file can set defaults for logging and locking. The CLI always takes precedence.

Default location: ~/.config/learningfoundry/config.yml

logging:
  level: INFO      # DEBUG | INFO | WARNING | ERROR
  output: stdout   # stdout | stderr

locking:
  sequential: false          # default for all curricula on this machine
  lesson_sequential: false

Pass a custom config location with -c / --config.


Development Setup

Prerequisites

  • Python 3.12+
  • pyve (virtual env manager used in this project)
  • pnpm 9+ and Node.js 18+

Setup

git clone https://github.com/pointmatic/learningfoundry.git
cd learningfoundry

# Create the Python environment and install the package in editable mode
pyve init
pip install -e .

# Create the test runner environment and install dev dependencies
pyve testenv --init
pyve testenv --install -r requirements-dev.txt

Running Tests

# Fast unit + integration tests (~2 min)
pyve test

# End-to-end SvelteKit smoke tests (requires pnpm, ~15 s extra)
pyve test tests/test_smoke_sveltekit.py -v

Linting and Type Checking

pyve testenv run ruff check .
pyve testenv run mypy src/

Project Structure

learningfoundry/
├── src/learningfoundry/
│   ├── cli.py              # Click CLI entry point
│   ├── config.py           # Configuration loading
│   ├── exceptions.py       # Exception hierarchy
│   ├── generator.py        # SvelteKit project generator
│   ├── integrations/       # Quiz / exercise / visualization providers
│   ├── logging_config.py   # Logging setup
│   ├── parser.py           # YAML parser + version dispatch
│   ├── pipeline.py         # run_build / run_validate / run_preview
│   ├── resolver.py         # Content reference resolver
│   └── schema_v1.py        # Pydantic v1 curriculum schema
├── sveltekit_template/     # SvelteKit app template (copied on build)
├── tests/                  # pytest test suite
├── requirements-dev.txt    # Dev dependencies
└── pyproject.toml          # Build config, ruff, mypy, pytest settings

License

Apache 2.0 — see LICENSE.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

learningfoundry-0.61.0.tar.gz (145.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

learningfoundry-0.61.0-py3-none-any.whl (166.7 kB view details)

Uploaded Python 3

File details

Details for the file learningfoundry-0.61.0.tar.gz.

File metadata

  • Download URL: learningfoundry-0.61.0.tar.gz
  • Upload date:
  • Size: 145.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for learningfoundry-0.61.0.tar.gz
Algorithm Hash digest
SHA256 16bcca2541c43c8e4de16e4dce4b3951bfce1ad87d9078253536b990bd914ae5
MD5 6b3f4c9da358a0ec716a09c17cb932d7
BLAKE2b-256 703bfcf51e5c86cf5c6f1a8102d7881c1fb6252e6527af1f3df0c7c71b94883d

See more details on using hashes here.

Provenance

The following attestation bundles were made for learningfoundry-0.61.0.tar.gz:

Publisher: publish.yml on pointmatic/learningfoundry

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file learningfoundry-0.61.0-py3-none-any.whl.

File metadata

File hashes

Hashes for learningfoundry-0.61.0-py3-none-any.whl
Algorithm Hash digest
SHA256 6d93abdf0dc6b1e8d7c870e8f031e6dac064f916873849f1818d73cce6c68c93
MD5 8415326ba3f4603ee6b6b5c9ae36cbab
BLAKE2b-256 e6e6c4829e0c9f8d01cf088cb4313a198070566e838c3daa46287e30f72634a2

See more details on using hashes here.

Provenance

The following attestation bundles were made for learningfoundry-0.61.0-py3-none-any.whl:

Publisher: publish.yml on pointmatic/learningfoundry

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page