SSEC-JHU autora_gui

autora-gui is a browser-based, visual workflow editor for AutoRA (Automated Research Assistant). It lets you assemble AutoRA theorists, experimentalists, and experiment runners into a closed-loop research workflow by dragging components onto a canvas, wiring them together, and editing their parameters — then export the result to a runnable AutoRA workflow.

The editor is served by a FastAPI backend that reads component and schema definitions from autora_gui/JSON/, with a React front-end (autora_gui/react_app). A live, self-contained build is deployed to GitHub Pages:

🔗 Live demo: https://ssec-jhu.github.io/autoRA-gui

Repository layout

Quickstart Guide

# 1. Clone the repository
git clone git@github.com:ssec-jhu/autoRA-gui.git
cd autoRA-gui

# 2. Create and activate a conda environment
conda create -n autora_gui python=3.11 pip
conda activate autora_gui

# 3. Install the package (editable) with dev dependencies
pip install -e ".[dev]"

# 4. Start the FastAPI backend (from the react_app directory)
cd autora_gui/react_app
uvicorn server:app --reload --port 8000

# 5. In a second terminal, start the React dev server
cd autora_gui/react_app
npm install
npm run dev

The editor is then available at http://localhost:3000 (the front-end talks to the backend on port 8000). Prefer no local setup at all? Use the live demo.

Installation

Create a conda virtual environment

For additional commands see the Conda cheat-sheet.

• Download and install either miniconda or anaconda.
• Create a new environment: conda create -n autora_gui python=3.11 pip (Python 3.11+ is required).
• Activate it: conda activate autora_gui.
• cd into the repo directory.
• Install the Python dependencies. There are two common options:
  • For development with tox (recommended): pip install -r requirements/dev.txt.
  • To run everything outside of tox: pip install -r requirements/all.txt.

Install the front-end dependencies (Node.js)

The React front-end requires Node.js (v20 recommended):

cd autora_gui/react_app
npm install

Build with Python

• cd into the repo directory and activate your environment: conda activate autora_gui.
• Build and install the package: pip install ..
• For dev/editable mode (repo changes are reflected in the installed package on kernel restart — the preferred method for development): pip install -e .. NOTE: to pull in the looser-constrained dev extras at the same time, use pip install -e ".[dev]".
• To build the distributable wheel and source tarball (this is what CI publishes to PyPI): tox -e build-dist (or python -m build). Artifacts are written to dist/.

Run with Python

The editor has two processes: the FastAPI backend and the React front-end.

• Activate your environment and cd into the front-end directory:

  conda activate autora_gui
  cd autora_gui/react_app
  

• Start the backend API:

  uvicorn server:app --reload --port 8000
  

• In a second terminal, start the front-end dev server:

  cd autora_gui/react_app
  npm run dev
  

• Open http://localhost:3000 in your browser.

To produce the single-file, self-contained build (the one deployed to GitHub Pages), run python build_standalone.py from autora_gui/react_app and open the generated index.html.

Docker

The repository ships a Dockerfile that installs the production dependencies (requirements/prd.txt) and serves the FastAPI backend. On every push, CI builds this image and publishes it to the GitHub Container Registry at ghcr.io/ssec-jhu/autora-gui (see ci.yml).

Build

• Download & install Docker — see the Docker install docs.
• cd into the repo directory.
• Build the image: docker build -t autora-gui ..

Run

• Run a container, mapping the backend port:

  docker run -d -p 8000:8000 autora-gui
  

  The backend API is then available at http://localhost:8000.
• Alternatively, pull the pre-built image from the registry, e.g.:

  docker pull ghcr.io/ssec-jhu/autora-gui:main
  

The container serves the backend API only; the React front-end is built and served separately (see Run with Python or the live demo).

Testing

NOTE: The following steps require pip install -r requirements/dev.txt.

Using tox

Tox runs each check in its own isolated virtual environment, matching the CI on GitHub Actions (see ci.yml).

• Run the full suite (style, security, tests, docs, and package build): tox.

• Run an individual environment with tox -e {env}:

  Command	What it does
  tox -e check-style	Lint and format check with ruff.
  tox -e check-security	Security scan with bandit.
  tox -e format	Auto-format code and sort imports with ruff.
  tox -e test	Run the pytest suite with coverage.
  tox -e build-docs	Build the Sphinx HTML docs.
  tox -e build-dist	Build the wheel and source distribution.

Outside of tox

The following assume all requirements are installed into your conda environment, e.g. with pip install -r requirements/all.txt.

NOTE: Tox will run these for you; the steps below are for running them directly.

Linting

Checks style, imports, and simple code-analysis issues using ruff.

• cd into the repo directory and activate your environment.
• Check formatting: ruff format . --check.
• Check lint rules: ruff check ..
• These can run automatically on every git push by installing the provided pre-push git hook: cp ./githooks/pre-push .git/hooks/; chmod +x .git/hooks/pre-push.

Security Checks

Checks for security concerns using Bandit.

• cd into the repo directory.
• Run: bandit -c pyproject.toml --severity-level=medium -r autora_gui.

Unit Tests

Tests core package functionality at a modular level with pytest. Test suites live under autora_gui/tests, autora_gui/js_app/tests, and autora_gui/react_app/tests.

• cd into the repo directory.
• Run all Python tests with coverage: pytest --cov=./.
• Run a specific test: pytest autora_gui/tests/test_util.py.
• Run the React front-end tests: from autora_gui/react_app run npm test (or npm run test:coverage).

Build Docs

Builds, tests, and lets you view the Sphinx documentation.

• The simplest route is tox -e build-docs.
• To build manually:
  • pip install -r requirements/docs.txt.
  • cd docs.
  • make clean.
  • make html.
  • View the docs in your browser: open docs/_build/html/index.html.

The docs are automatically built and deployed to https://ssec-jhu.github.io/autoRA-gui.