Editor Python and R setup

Rime Editor uses the same interpreter model as the CLI. Python and R nodes run through configured local interpreters; SQL and JavaScript run through the built-in runtime paths.

Resolution Order

When the editor runs a DAG, Python and R interpreter paths resolve in this order:

1. project/editor selection
2. DAG interpreters: block
3. environment variables such as RIME_PYTHON_BIN and RIME_RSCRIPT_BIN
4. system fallback (python3, Rscript)

The important rule is that interpreter choice is execution configuration, not pipeline structure. The DAG can declare defaults, but local machines and CI can override them.

specification_version: "2.1"

interpreters:
  python: .venv/bin/python
  r: /usr/local/bin/Rscript

What The Editor Should Show

The environment surface should stay lightweight:

• selected Python path
• selected R path
• automatic status after a path changes or before a DAG run
• missing package messages when a script node cannot run
• raw interpreter output only when something fails

There should not be a separate “probe” ritual for normal use. If the user tries to run the DAG with default paths, the editor can check automatically and surface the result in context.

Script Nodes

Language nodes declare named data slots in YAML:

- id: features
  kind: python
  source: scripts/features.py
  in:
    cohort: repeat_visitors
    threshold: params.threshold

The editor should render cohort as an input edge and keep scalar params.* bindings visible in the spec/config surface.

CLI Equivalent

export RIME_PYTHON_BIN=.venv/bin/python
export RIME_RSCRIPT_BIN=/usr/local/bin/Rscript
rime run pipeline.dag.yaml

For runtime behavior, see Polyglot runtime.