Skip to content

Latest commit

 

History

History
53 lines (37 loc) · 5.86 KB

README.md

File metadata and controls

53 lines (37 loc) · 5.86 KB

missing-elements.github.io

The site of the missing-elements web components, at https://missing-elements.github.io: the docs of the PDF viewer, and the pages that make H5P content with an AI chat and the h5p-author skill.

What it serves

Path What it is Source
/ The project list, the tools (the build page first, with its how-to), and a footer of links src/index.njk; the lists are in src/_data/site.js
/pdfjs-viewer/ The docs of <pdfjs-viewer-element>, with a live preview src/pdfjs-viewer/
/h5p/build/ The build page, for people: what it makes, a sample that plays, and the build of a chat's source into a .h5p, in the visitor's browser src/h5p/build/
/h5p/create-with-ai/ The how-to, for people: installing the skill, asking a chat, the prompts to start from, the types, and what to do when something goes wrong src/h5p/create-with-ai/
/h5p/skill/ The skill as zips to install: h5p-author.zip for Claude's Skills and for Grok, attached to a message, h5p-author-chatgpt.zip for ChatGPT's Plugins; and a page that lists them and the bundles' addresses, each in full, which no page links to src/h5p/skill/; the zips written by vite.config.ts
/h5p/bundles/ The skill's build tool and a library bundle per content type, for a chat that runs Node but cannot reach npm; and h5p-author-build.tar.gz, the tool and every library in one archive, for a person to attach to a chat whose code can download nothing written by vite.config.ts
/h5p/build/types/ The skill's type files, which the build page copies for the chat written by vite.config.ts

A chat takes the skill installed: from a zip in Claude and ChatGPT, attached to a message in Grok, which has no place to install a skill, or with npx skills add in an agent with a terminal, as /h5p/create-with-ai/ says. Where the skill cannot reach npm, it builds from the tool and a bundle in /h5p/bundles/, at the addresses /h5p/skill/ lists, if it can run Node and fetch a file: Grok did, on 2026-11-13. Where its code runs Node but can fetch nothing, as in ChatGPT, it builds from h5p-author-build.tar.gz once the person attaches it or keeps it in a ChatGPT Project (the owner's free account, 2026-11-14), and otherwise hands them the source for /h5p/build/, which builds in their browser. The tool is a Node.js program, so a chat whose code runs no Node, as most chat apps for local models, hands over the source; an agent with a terminal and a local model builds offline from the archive.

Until 2026-11-13 the site's prompts sent a chat to read the skill at /h5p/skill/ instead. That worked only in chats that now take the zip, so the prompts name the installed skill, and no page links to /h5p/skill/. Its page stays, for the skill's route 3 and the copies already installed, which take the bundles' addresses from it. The skill's text files were served there one by one too until h5p-author 0.10.0: the zips carry them, so the site serves them only as zips. A copy of 0.9.7 or older, in a chat that has its text alone, and a prompt copied before 2026-11-13 find no files there now.

Develop

Needs Node 22.12 or later and pnpm.

pnpm install
pnpm dev        # Eleventy's server, http://localhost:8080
pnpm build      # type-checks, then Eleventy writes _site/, then Vite writes /h5p/ into it
pnpm preview    # serves _site/ as GitHub Pages does, http://localhost:4173

pnpm dev shows the pages as Eleventy writes them: the build page's script and its sample are there only after a pnpm build.

How it is built and deployed

  • Eleventy writes the pages from src/ into _site/. They ship no framework. Their scripts are the theme toggle and the code-block buttons (src/assets/site.js), the pdfjs-viewer live preview (src/pdfjs-viewer/docs.js) and the build page.
  • Vite then writes /h5p/ into the same _site/: the build page's script, bundled with the H5P player, its runtime and the library pack; the build page's sample; the type files; and the bundles. That is build output: git holds none of it, and the bundles alone are over 100 MB.
  • A push to main deploys through pages.yml. A pull request is built by build.yml.

What the site pins

The pages run published releases, pinned exactly in package.json, and a release moves what the site serves:

Package What it moves
@missing-elements/pdfjs-viewer The docs' version numbers, and the viewer in the live preview
@missing-elements/h5p-author The build the build page runs, its sample, the type files, the skill's zips at /h5p/skill/, and the tool in /h5p/bundles/
@missing-elements/h5p-libraries The library pack: the build page's libraries, and the bundles
@missing-elements/h5p-offline-player, @missing-elements/h5p-runtime The player on the build page, and its runtime

To move a pin, change its version in package.json, run pnpm install, then pnpm build, and commit the lockfile with it. Dependabot opens the pull requests for the pins as their own, apart from its group for everything else, and takes a new @missing-elements release the day it comes out where it holds other packages back for three days: .github/dependabot.yml and pnpm-workspace.yaml say why.

The rules behind the H5P pages, and what was measured to reach them, are in design/author.md of h5p-toolkit, where the skill and its packages live.