How to access the internal graph representation (AST?) of a parsed flowchart diagram?

[digeomel]

Hello, I would like to get access to the internal Mermaid representation of a parsed flowchart diagram in order to know e.g. which nodes are before and which nodes are after a particular node. Then I would be able to dynamically construct a new textual representation of a subset of that diagram and pass it to Mermaid for rendering again.

I noticed in the docs that there used to be a mermaid.mermaidApi object which now seems to be deprecated.

I also tried with the @mermaid-js/parser parse method, but it seems that it does not support diagrams of type flowchart.

So, is there a way to do it?

[thomas-jakemeyn]

I also need a programmatic access to a erDiagram and I am facing the same problems. parse() does not return the diagram and can only be used for validating the syntax. On the other side,@mermaid-js/parser does not support ER diagrams.

My workaround so far has been to patch the mermaid library (using patch-package) in order to add diagram to the object that is returned by parse() (here). That way, I can access the diagram that has been parsed based on the official Jison grammar and I don't need to build my own grammar using Langium, ANTLR, Tree-sitter, etc.

Is there any chance for that change to be integrated into the library? If not, what are the alternatives?

[digeomel]

@thomas-jakemeyn was thinking of doing the same, but I thought I'd try here first. I don't understand why they limited the interfaces they return in the parser, since they need to support all of them internally anyway.