markdown-confluence
Support for plantuml/graphviz or general preprocessing of file
Hi,
This has worked really well for me on first use but I've got some issues with where I've been using https://github.com/joethei/obsidian-plantuml and https://github.com/QAMichaelPeng/obsidian-graphviz extensively in my documents.
I've preferred these to Mermaid for various reasons so it would be good to think about how to support these other plugins or indeed a more general approach for it to publish processed Markdown after other plugins have added stuff (e.g. as would be the case for plugins like https://github.com/obsidian-tasks-group/obsidian-tasks that generate content).
I'm happy to have a try of this myself this week in my own fork but I thought I'd raise an issue first to see if this has been thought of and if there's opinions on how to approach things.
I'm thinking either:
- There's a way to configure a preprocesser command that takes the raw Markdown and replaces various codeblocks with images. Then that Markdown is sent for Confluence conversion, not the original. I already publish one vault as HTML via https://docs.getpelican.com/en/4.8.0/index.html so I've done this approach before to preprocess Markdown. The bit that I'd need is a clear hook to get markdown-confluence to call out to first.
- I can literally copy-paste the Mermaid plugin in this repo and make a sibling version for PlantUML.
- I'm not experienced with writing Obsidian plugins but maybe there's a way for this one to feed off the output of the others I use?
I don't think my company has PlantUML or Graphviz support set up on Confluence so a path of least resistance for me would be to flatten to images before this plugin runs and leverage the fact it supports image uploads.
Would love to hear some thoughts/input and I'll have a try at what I can get working.
Hey,
I have had PlantUML support requested from a friend. I extracted the logic so it should be "easy" to copy the Mermaid to support a PlantUML one. https://github.com/markdown-confluence/markdown-confluence/blob/main/packages/lib/src/ADFProcessingPlugins/MermaidRendererPlugin.ts The current plan in my head was to offer a PlantUML plugin that called out to the PlantUML website by default to render to images with a configuration option that would allow the user to define their own for privacy reasons.
I hadn't considered graphvis but am open to adding a Plugin for that. It would need to have a way to detect if graphvis is installed and if not the bypass the rendering.
I haven't looked at how to hook into the rendered Markdown in Obsidian. Currently the logic calls vault.cachedRead https://github.com/markdown-confluence/markdown-confluence/blob/main/packages/obsidian/src/adaptors/obsidian.ts#L67 I am not aware of a method in the current API surface to get additional MD from other plugins. If there is a way happy to have a look at updating the code to use that path.
Long term I want to make it possible for users to dynamically load their own Plugins from outside my codebase but that hasn't been a priority yet.
Nothing you have described sounds very complex it will take a bit of testing and adding some new options that need to be explored for the various usages of the code.
Happy to accept contributions :) If you want to chat though the code as you are working on it feel free to join Discord and we can continue the conversation there.
I look forward to having it as well. Unfortunately, I'm not knowledgeable in JS/TS to contribute.
