Skip to main content

Reproducible code (Reproduce)

The Reproduce view (in the Lab zone) shows the exact, re-runnable code behind your results — a deterministic Python script that replays the engine calls your session ran, in order. It is not AI-generated: each call is rendered into real Python (statsmodels / scipy / lifelines / …) from the engine's per-method templates, so it runs outside Statios with only open-source dependencies. Drop it into a paper's code appendix or a "code availability" statement, or re-run it to verify your numbers.

Generate the script

  1. Run some analysis — chat an analysis, run a Guided plan, or run any test.
  2. Switch to the Lab zone (Ctrl+4) and select the Reproduce sub-tab.
  3. The script appears with a provenance header (dataset, Statios version, how to run).

Choose what to include

Use the source selector at the top:

  • Full session — everything you ran, grouped by source.
  • Guided — the compiled Guided plan (planned steps; gate decisions are shown as comments).
  • Log — the analyses you ran from chat, with your prompts as comments.
  • Outputs — the standalone results in the Output zone.

Steps that could not run (a not-yet-implemented method, or a node with unbound parameters) are listed in a trailing # NOTE: block rather than emitted as calls, so the script always stays valid and faithful to what actually executed.

Copy / download

  • Copy puts the script on your clipboard.
  • Download .py saves it as a file.

Cite the underlying methods via the report's APA reference list, and cite Statios itself as software (see the report's references).