robots-from-jupyter / robotlab

Experiments in building installers for JupyterLab, Robot Framework and Friends
BSD 3-Clause "New" or "Revised" License
25 stars 3 forks source link

RobotLab Build Status

A one-click installer for Robot Framework, JupyterLab, Firefox, and Friends

Get RobotLab

Start the download now for your platform:

Linux | MacOSX | Windows

... or view the release notes.

Install RobotLab

You'll want at least 2GB available to install and run RobotLab

Remember where you install RobotLab!

We recommend someplace

  • short, e.g. C:\robotlab or /home/myuser/robotlab
  • with no spaces in the path, e.g. c:\Robot Lab
  • not shared on something like Dropbox, or a network drive

Run RobotLab

During installation, RobotLab will try to create a desktop icon:

There is an icon on my desktop/menu

Click it to launch the RobotLab server in a console window. It will also attempt to launch the bundled Firefox. The application log will keep running in the background, and sometimes useful information will appear there.

Nope, no icon here

If you don't see an icon, launch a console, e.g. bash on Linux/MacOSX or cmd.exe on Windows.

Linux/MacOSX Windows
start a prompt open your terminal or console application open cmd.exe from the start menu
activate the robotlab environment type source /path/to/robolab/bin/activate
press ENTER
type c:\path\to\robotlab\Scripts\activate
press ENTER
start robotlab type robotlab
press ENTER
type robotlab
press ENTER

Still didn't work

If you're still stuck, open an issue. There should be some template information to provide.

Advanced RobotLab Commands and Invocations

# assuming you've already `activate`d
robotlab [.]               # basically just jupyterlab
robotlab-tutorial [.]      # copy the robotkernel tutorial to this directory
robotlab-examples [.]      # "    "   "           examples "  "    "
robotlab-shortcuts

Uninstall RobotLab

Motivation

While Robot Framework has no dependencies beyond the Python standard library, using it for non-trivial testing or process automation usually requires a fair number of additional Python dependencies, and even some more exotic ones. We wanted to run a workshop for RoboCon 2019 and show interactive editing/running with robotkernel.

The focus of this workshop, running Robot tests and tasks interactively, required the Jupyter stack. To demonstrate some of its more advanced features, some fairly extensive extra libraries from both the Robot Framework and the scientific Python ecosystems were needed, and getting these to install smoothly on every attendee's computers can prove to be... interesting.

To make a three-hour workshop reasonable, we needed a distribution approach that was:

Field Notes

At our 2019 workshop, graciously hosted by Siili, (whose network was super fast!) our room full of attendees, on three distributions of Linux, two versions of MacOSX and two versions of Windows were all up and running in under 30 minutes. We got through all of the tutorial before lunch!

Tools

RobotLab is built with:

conda packaging

Once it has matured, robotkernel will be available from conda-forge, the community-driven upstream of the Anaconda Distribution, which is also built with constructor.

So for the RobotLab installers, and supporting examples and tutorials, we need to build a number of dependencies.

robotlab

JupyterLab's build chain has some negative externalities for end users, namely an install- or run-time dependency on NodeJS and npmjs.org when using any labextensions other than the built-in set (e.g. Notebook, Terminal, Console, Editor, etc.). Because, for the purposes of the workshop, we want to get to the Good Stuff of running Robot notebooks and not spend a bunch of time debugging nodejs and webpack, we've added a few choice JupyterLab extensions:

All of these are built outside of the conda-build cycle.

This asset, and all our dependencies, are wrapped into a conda package which exposes some useful commands, which can do most of the things jupyter lab can do, as well as create desktop shortcuts, and unpack the tutorial content.

robotlab works like jupyter lab, while robotlab-extension works like jupyter labextension. This isn't a toy installation: with the bundled nodejs, an intrepid user can still install any of the labextensions that are compatible with the version robotlab was built with: as of writing, 1.1.x.

constructor

All of the dependencies are captured in construct.yaml.in. In addition to required dependencies, a number of extra libraries are included to showcase some of the features of using Robot Framework interactively.

On Browsers and WebDrivers

RobotLab includes Mozilla Firefox and geckodriver, with versions that will be supported for years, not weeks, by a team that is committed to open source and web standards.

webdriver for Microsoft Internet Explorer

It's pretty easy to get webdriver for Microsoft Edge, but...

chromedriver for Google Chrome

It's also pretty easy to get chromedriver for Google Chrome, but...

Work On RobotLab

Contributions are welcome! See the [issues][] for outstanding work we've been discussing.

From the command line e.g. bash or cmd.exe or the Anaconda Prompt, run:

anaconda-project run build:lab
anaconda-project run build

Test Robotlab

anaconda-project run test

Checking RobotLab

anaconda-project run lint
anaconda-project run integrity