IRIS DocBook -> Markdown

This is a browser extension that converts online InterSystems IRIS DocBook documentation pages into clean, portable Markdown with an offline conversion pipeline and a safe preview, making it easy to incorporate product-specific information into modern AI assisted workflows.

By turning a live documentation page into a single Markdown artefact you can paste into ChatGPT, a custom GPT, or an agentic development tool, you can supply relevant project context. Context like APIs, configuration steps, ObjectScript examples, and system-level constraints can all help reduce hallucination and improves code generation quality.

The output is designed to be LLM friendly by flattening tabbed content, normalizing code blocks with language hints, and promoting glossary entries and headings for better retrieval. Optionally, you can strip links or tables when you need to save tokens.

What it does

When you click the extension icon on a supported DocBook page, it:

• Extracts the main article DOM from the active tab (on-demand).
• Converts DocBook HTML to Markdown using a custom pipeline.
• Opens a viewer tab with:
  • Markdown output (editable textarea).
  • Sanitized preview (sandboxed iframe).
  • One-click copy/download.
  • Optional link/table stripping.

Supported pages

This extension is intentionally scoped to InterSystems IRIS documentation:

• Host must be: docs.intersystems.com.
• Path must include one of:
  • /csp/docbook/DocBook.UI.Page.cls.
  • /csp/docbook/Doc.View.cls.
• URL must include a KEY=... query parameter.

If you click the icon on an unsupported page, the viewer opens with an explanatory error.

Quickstart

If you just want to use the extension, you can load the pre-built dist/ bundle from this repo.

Download the repo as a ZIP

• On GitHub, click Code -> Download ZIP.
• Unzip it locally.

You should have a folder that contains a dist/ directory.

Load into Chrome

1. Open chrome://extensions.
2. Enable Developer mode.
3. Click Load unpacked.
4. Select the dist/ folder from the unzipped repo.

Load into Edge

1. Open edge://extensions.
2. Enable Developer mode.
3. Click Load unpacked.
4. Select the dist/ folder from the unzipped repo.

Usage

Open a supported IRIS DocBook page:

• Click the extension icon.
• A new tab opens showing the converted Markdown.

In the viewer you can:

• Copy Markdown
• Download .md
• Copy Source URL
• Clear Session Data
• Toggle:
  • Strip Links (Keep Text)
  • Remove HTML Tables

Output notes

The conversion pipeline does a set of targeted normalizations aimed at DocBook pages:

• Flattens tabbed blocks into headings + content.
• Promotes table titles and glossary entries into headings.
• Normalizes code blocks and attempts language detection (e.g., ObjectScript/SQL/XML/JS/Python).
• Converts admonitions (note/important/caution/tip/warning) into blockquotes with a label.
• Removes images (replaces with a text placeholder when an alt value exists).
• Repairs some malformed list structures.
• Emits tables as raw HTML by default (optionally removable in the viewer).

Markdown conversion is done via Turndown with a few overrides for better fidelity.

Security & privacy

This extension is built to be conservative:

• No network access from extension pages:
  • Extension CSP includes connect-src 'none'.
• Runs only when you click the extension icon.
• Only extracts DOM from the active tab, and only on supported DocBook URLs.
• Extracted content is stored in a session-scoped manner using chrome.storage.session.
  • You can clear it immediately via Clear Session Data.
• Preview is hardened:
  • Markdown is rendered to HTML locally.
  • HTML is sanitized with DOMPurify and a tight allow-list.
  • Links are rewritten to allow only http(s) URLs on docs.intersystems.com.
  • Preview is displayed in a sandboxed iframe using srcdoc.
  • Scripts are blocked.

Development

If you want to modify the extension or build dist/ yourself, you’ll need Node.js.

Prerequisites

• Node.js (recent LTS recommended)

Install dependencies

npm install

Run a local dev server

npm run dev

Build the unpacked extension

npm run build

Then load the generated dist/ folder via Load unpacked (see Quickstart).

Project structure (high level)

• src/background.ts - action click handler; validates URL; extracts page DOM; stores session payload; opens viewer.
• src/viewer.ts - reads session payload; converts to Markdown; renders output + safe preview; copy/download actions.
• src/convert.ts - conversion orchestrator (DOM passes + Turndown).
• src/domPasses.ts - targeted DOM normalization passes for InterSystems DocBook HTML.
• viewer.html - viewer UI (Markdown + Preview tabs).
• icons/ - extension icons.