jupyterlab dyno (aka Dyno Lab)

A JupyterLab extension for DSGE modeling.

Installation

• With micromamba:

micromamba install -c https://repo.prefix.dev/econforge jupyterlab_dyno

• With pixi:

pixi add jupyterlab_dyno

A conda-forge release of jupyterlab_dyno is in the works so the above command should work for you. If it doesn't then add the "https://repo.prefix.dev/econforge" to the list of channels in your pixi project and try to run it again.

Development

Requirements

• Pixi: a reproducible package management tool

Development Installation

pixi install

Run JupyterLab

pixi run lab

Live development (watch + lab)

pixi run watch

This runs jlpm watch and jupyter lab in parallel. Every saved change will be rebuilt automatically — just refresh JupyterLab to load it.

Dyno Options Sidebar

Opening a supported model file (e.g. a .mod or .dyno file) now also exposes a sidebar panel named "Dyno Options" on the left side of JupyterLab.

Use it to tweak per-file run parameters without editing global settings:

• Order: numerical order (default 1)
• Steady state only: if checked, only the steady state is computed
• Preserve scroll: toggle preservation of the output panel scroll position across re-renders for the current file

Changing any option automatically re-renders the currently active Dyno view. Each file remembers its own option set for the session; switching between open model files updates the panel to reflect that file's stored values.

These per-file options are merged on top of the extension's global settings (defined in Settings -> Advanced Settings Editor -> Dyno Lab).

Important files

• pixi.toml contains the development environment dependencies and scripts
• src/index.ts contains all of the extension's code and logic
• schema/plugin.json defines the user-facing extension settings
• package.json contains the core JupyterLab nodejs dependencies and their versions
• recipe.yaml contains a conda-build recipe for the extension

Packaging and release

For conda-forge

A conda-build / rattler-build v1 recipe can be found in the recipe folder. You can test it out by installing rattler-build if you don't have it already with

pixi global install rattler-build

and then by running the build command:

rattler-build build -r recipe

The version that is built with this command is not the version available locally, but rather a github release that can be modified by publishing a new release on the EconForge repo and then editing the context of the recipe.yaml file to reflect the new version tag.

For a prefix.dev channel

If you just want a .conda file in order to publish the extension on a custom conda channel (like https://prefix.dev/channels/econforge), this recipe is not needed, and you only need to run the following command in the project root.

pixi build