Feat: Sequence Diagram SVGs into Mermaid renders

[hanzalahwaheed]

Fixes #792

I've tried to convert the diagrams into mermaid renders. Here are the tow outputs that I have. Need help choosing the better one.

serial coupling diagram

reference: https://precice.org/couple-your-code-coupling-flow.html

original screenshot

mermaid render A

This is a more direct one-to-one mapping of the diagram.

mermaid render B

This is a more semantic render.

@MakisH which one do you think is a better render? I'm inclined towards A but the loop box of B is also nice.

[MakisH]

Thank you @hanzalahwaheed, this is useful. We will eventually review it and see how it might work together with other documentation as well.

[precice-bot]

This pull request has been mentioned on preCICE Forum on Discourse. There might be relevant details there:

https://precice.discourse.group/t/gsoc-2026-introduction-hanzalah-waheed/2745/5

[MakisH]

Overall, I think this is nice for new graphs, but I am not yet convinced of the value that the Mermaid version of this figure would add. I would suggest not investing more work on this at the moment, let's just use this PR to evaluate the option.

Some comments, in arbitrary order:

• The SGV (made with TikZ) also has some colors, which point to the two participants. Can we color boxes or text differently?
• In "Approach A", the two participants seem like they are taking turns in calling initialize(). This is not correct: up to that point, they are decoupled. Similarly, in "Approach B", the calls to initialize() and finalize() seem to be one after another.
• In "Approach B", I like the loop box.
• In the SVG, the indentation relates to the different software layers, and it is nice to have.
• An advantage of the SVG version is that it can easily be exported and used in a PDF document. How would we integrate Mermaid graphs there?

I tend towards not migrating this figure to Mermaid, but simpler/new figures would be fine to have directly in Mermaid. @fsimonis what do you think?

[MakisH]

Closing due to inactivity, and since this is far from a state to get merged.

In any case, I think this is a nice proof-of-concept, and thank you for contributing! The main technical work in #726 was already very useful.