Graph Visualization

After compiling a pipeline, describe_graph() returns a Mermaid diagram string that shows every node and edge in the compiled graph. Paste it into GitHub Markdown, documentation, or mermaid.live for a visual DAG.

describe_graph()

from neograph import compile, describe_graph

graph = compile(pipeline)
print(describe_graph(graph))

Output (varies by pipeline):

%%{init: {'flowchart': {'curve': 'linear'}}}%%
graph TD;
  __start__([<p>__start__</p>])
  decompose(decompose)
  classify(classify)
  report(report)
  __end__([<p>__end__</p>])
  __start__ --> decompose;
  decompose --> classify;
  classify --> report;
  report --> __end__;

Where to paste it

GitHub — wrap in a ```mermaid fenced code block in any Markdown file, PR description, or issue.
mermaid.live — paste the raw string into the editor for an interactive diagram.
Documentation — Starlight, Docusaurus, and other doc frameworks render Mermaid natively.
CLI — print to stderr/stdout during development for a quick topology check.

Error handling

If the compiled graph does not support Mermaid rendering (e.g., a non-standard backend), describe_graph() returns "(graph visualization not available)" instead of raising.

Dev-mode DAG summary

Set the NEOGRAPH_DEV environment variable to get an automatic DAG summary printed to stderr after every compile() call:

NEOGRAPH_DEV=1 python my_pipeline.py

Output on stderr:

[neograph-dev] Compiled 'ingestion' (4 nodes):
  START -> decompose
  decompose -> classify
  classify -> report
  report -> END

This shows every edge in the compiled graph, including conditional edges (marked [conditional]). The summary prints automatically — no code changes required.