Exporting a Case

FlowGraph offers three complementary ways to get data out of the editor, depending on whether you want to persist your work, reuse a sub-graph, or produce a runnable Kratos case.

1. Save / Load the full graph

Save downloads the complete editor state as graph.json and Load restores it. This is the FlowGraph-native format: it preserves nodes, widget values, positions and links exactly, but it is not a Kratos case. Use it to checkpoint work-in-progress. See The Toolbar.

2. Export / Import a selection

Export writes only the selected nodes and their internal links to selection.json; Import loads such a file back, remapping IDs so it never clashes with existing nodes. Use selections to build a personal library of pre-wired blocks (a configured solver stack, a standard set of boundary conditions, …). See The Toolbar.

3. Export a runnable Kratos case

To produce something Kratos can actually run, use an Export case files node (IO/DownloadProblem), implemented in public/js/nodes/IO/export_case.js.

Connect the output of your top-level analysis/orchestrator graph into the node's data input, then trigger its Download button. The node uses JSZip to bundle:

• ProjectParameters.json — the generated Kratos configuration, and
• each material JSON file referenced by the graph.

into a single case.zip.

Steps

1. Build and wire your graph (see Getting Started).
2. Add an Export case files node and connect the final analysis output into it.
3. Click Generate to compute the JSON.
4. Press the node's Download button to save case.zip.
5. Unzip it into your Kratos working directory and run as usual (e.g. python MainKratos.py).

Running from the backend (optional)

If FlowGraph is running on a machine with a local Kratos build, the backend also exposes a run_simulation route that writes ProjectParameters.json into the configured working_dir and launches MainKratos.py, streaming the output back. This requires the kratos_root, working_dir and python_binary settings described in Configuration.