Rendering Plant UML

[PommeDeTerre0]

Hey everyone,

In my company we're big fans of PlantUML diagrams to describe requirements, currently we render these diagrams separately and then add them to the requirements as images.

I see that there is a backlog item for UML support, and I would like to have a go at it since I've implemented PlantUML rendering in some other projects in previous jobs.

I'm not very familiar with the strictdoc code-base though, and some help to point me in the direction of where such implementation would be appropriate would allow me to get started much faster !

[stanislaw]

Hi @PommeDeTerre0, thanks for asking! It would be great if you could implement the PlantUML support. I am, of course, biased but I think you should get around with the StrictDoc codebase and architecture pretty easily.

I would suggest that you follow how these two most similar features are implemented: MathJax and Mermaid. You should start by looking in the following place where the ProjectConfig file has feature toggles to activate these features. Look for:

    def is_activated_mathjax(self) -> bool:
        return ProjectFeature.MATHJAX in self.project_features

    def is_activated_mermaid(self) -> bool:
        return ProjectFeature.MERMAID in self.project_features

and see where these methods are used.

Is PlantUML implemented as a JS library that does all markup processing on the browser side or we also need to preprocess PlantUML syntax before it gets printed as HTML?

[PommeDeTerre0]

Hey !
Sorry for the late reply. Work got a bit crazy and the time I thought I could dedicate to this went away 😭

For me the way to go is to preprocess it, using either a local binary or a remote server, like it's done in plantuml-markdown.

But I heard you're already working to include python-markdown to strictdoc ? If that's the case it would probably be easier to ensure compatibility with python-markdown plugins

[stanislaw]

Yes, we are developing the Markdown support in StrictDoc but we are using this library:

    # Parsing Markdown in CommonMark mode. Converting Markdown to HTML.
    "markdown-it-py >= 4.0.0, == 4.*",

Do you think there is a way to make PlantUML work with markdown-it-py instead of python-markdown?

[PommeDeTerre0]

It's probably possible to work on a translation layer to use the plantuml-markdown extension with markdown-it-py, but it would be bit of a "hack". Maybe it would be easier to port one of the plantuml extension for JS mardown-it in python.

I don't have much time right now but this subject do interest me a lot, so I will try to explore this

[PommeDeTerre0]

Hi again,
I've started to look into it. I'm not used to jinja so it's a bit confusing at times, but I think I'm starting to understand how everything is structured.

I think the easiest way to do this would to, preprocess the plantuml description using either a local binary or the default public plantuml server and then simply display the image like it's done for classic images.

I still need to understand where exactly these operations need to be done

[stanislaw]

I have created a feature ticket out of this discussion. Let's continue there because PlantUML is something that we want to have as a working feature sooner or later.

#2791