Archi Forum

I'm finalizing Archi scripts for Confluence (https://github.com/yveszoundi/archi-confluence) publishing, via REST API integration: I originally had a working but less flexible version.

• A "convenient" dialog helps you store Confluence settings (URL, API token, etc.)
• After defining a publishing pipeline, you can update one or multiple pages on Confluence from Archi

Confluence Settings screenshot
settings_screenshot.png

Confluence Pipeline screenshot
pipeline_screenshot.png

Additional notes on top of the TLDR version.

Concepts

An archi-Confluence page is modeled as a Group element on a Sketch diagram:

• An archi-Confluence page contains either nested sections (Groups) or references to diagrams (Canvas, Sketch or default Archimate diagrams).
• Within an archi-Confluence page, the flow between sections is specified using connections. All page contents but the last element, must point once to something (either a section or diagram reference).

 
An archi-Confluence pipeline on a Sketch diagram contains one or multiple archi-Confluence pages:

• If you don't select anything on a sketch diagrams, all the archi-Confluence pages will be published.
• Alternatively, you can select specific archi-Confluence pages to publish.

The name of archi-Confluence page is used as Confluence page title (retrieve existing page information by title and Confluence space: page id, page version, etc.).

Confluence integration

Diagram documentation and images are published through the Confluence REST API (https://developer.atlassian.com/server/confluence/confluence-rest-api-examples/).

• You, create a Confluence API token (https://support.atlassian.com/atlassian-account/docs/manage-api-tokens-for-your-atlassian-account/) and keep it in your password manager.
• You, configure Confluence publishing settings in Archi: A utility script provides a dialog that helps persisting Confluence parameters in the Archi workbench preferences. This includes the Confluence username, API token, default space key, base URL.
• The script, retrieves Confluence page information, based on the page title: Page titles are unique in Confluence within a given space (space key).
• The script, renders any documentation associated to a view or 'section' using markdown (view/section documentation is parsed as Markdown and converted to HTML)
• The script, exports any referenced view/diagram image as PNG binary data
• The script, publishes text and image contents to Confluence pages accordingly to the desired sections/sub-sections.

Design decisions

It is more flexible to use Sketch diagrams than Canvas or regular Archimate views.

• You can connect directly 'view references' to each other without additional nesting.
• There's no need to think too much about the allowed structure: there are only groups, view references and connections (minimal validation in place).

I had a bit of time today to fix some of the typos in the code and test it.
I'm reasonably "satisfied" so far and things seem to be working as expected: I replaced my old Confluence workflows today with the newer scripts.

To aggregate content from an other thread, there are similar efforts from @Xavier Mayeur

• See https://github.com/xmayeur/archiScript and look for 'Export to Markdown.ajs'
• See https://github.com/xmayeur/toConfluence for the import/export script from/to Confluence API