Orchestra Visual Flow Programming

Orchestra VFP is a plugin for Python3 Jupyter Notebooks that enables data scientists and others less familiar with the Python Data Science toolchain to powerfully leverage the full range of data munging and (eventaully) machine learning tasks supported by Python itself.

Users more deeply familiar with Python code can use a hybrid workflow that "exports" python variables directly to the notebook in which Orchestra definitions are created.

Even as Orchestra offers greatly increased accessibility for data manipulation tasks, it does not force any compromises on those who wish to use the full power of Python.

Run Via Docker

• cd into the directory where you've cloned this repo
• docker build . -t orchestrajupyter
• docker run -p 8888:8888 orchestrajupyter
• You'll see a link printed to the command line like: http://localhost:8888/?token=f716322e4001692d4b6bb346b8e8d23b3af1db58780af32f
• Create a new Python3 Jupyter Notebook
• Click the Jupyter extension (see icon below) and go! (But be aware that this is just a "demo" environment -- your files will not be persisted after the container stops!)

Installation

Assuming you already have Jupyter and Python3 Installed locally, just run these commands:

jupyter nbextension install https://rawgit.com/strandedcity/orchestra-jupyter/master/dist/orchestra.js
jupyter nbextension install https://rawgit.com/strandedcity/orchestra-jupyter/master/dist/orchestra-libraries.js
jupyter nbextension enable orchestra

Expect to see this output:

# Downloading: https://rawgit.com/strandedcity/orchestra-jupyter/master/dist/orchestra.js -> /var/folders/sc/ps6q4zys24x3s4p711n8ltbw0000gp/T/tmps1m0cvrj/orchestra.js
# Copying: /var/folders/sc/ps6q4zys24x3s4p711n8ltbw0000gp/T/tmps1m0cvrj/orchestra.js -> /usr/local/share/jupyter/nbextensions/orchestra.js
# Enabling notebook extension orchestra...
#      - Validating: OK

And you're done! The next time you load a Jupyter Notebook, you'll have this icon available:

Click it, and you're off to the races.

Developers

Environment Setup

• The dev environment is very similar to the "production" environment, except you're using un-optimized versions of the code. To get started, clone this repo to a location of your choice (here, I assume your home directory), and install its "main" module as a Jupyter extension:

cd /path/to/containing/folder
git clone [email protected]:strandedcity/orchestra-jupyter.git
jupyter nbextension install /path/to/containing/folder/orchestra-jupyter --symlink
jupyter nbextension enable orchestra-jupyter/main

• Note that you cannot run both development and built copies of Orchestra side by side. So if you previously installed Orchestra via the instructions above, you'll need to:

jupyter nbextension disable orchestra

• Otherwise, you'll see this in your JavaScript console when starting Jupyter:

Uncaught SyntaxError: Identifier 'requireConfigFunction' has already been declared
    at orchestra.js?v=20171221222403:1

Building the JavaScript

• You will only need to build the JavaScript if you want to submit your code back to the main repo. To do so, you'll need a few npm dependencies:

npm install -g grunt-cli
cd /path/to/orchestra
npm install

# To build everything:
grunt

# To build only Orchestra (not its third-party dependencies):
grunt buildOrchestra

# To build only third-party libraries (not Orchestra code):
grunt buildLibraries

License

Orchestra Visual Flow Programming is Copyright 2017 Phil Seaton. https://www.phil-seaton.com.

Please note the license for Orchestra before you use or modify it. Orchestra is Licensed under the Creative Commons Attribution-NonCommercial 4.0 International License. Orchestra is free to use for non-commercial projects only, and modifications using Orchestra source may not be sold under any circumstances. Commercial licenses of Orchestra can be purchased for a reasonable fee, and may be offered in exchange for contribution to the Orchestra project. Please visit https://orchestravisualprogramming.com for more information.

Orchestra makes use of several third-party libraries, all of which are open-source projects under the MIT License or similarly permissive open-source licenses that do not restrict usage or modification:

• ThreeJS
• Handsontable Community Edition
• Numbro
• Pikaday
• Moment
• Backbone
• Underscore
• Bootstrap3-Typeahead (Apache 2.0 License)

The following are also dependencies, but are included already in Jupyter: JQuery, Bootstrap, RequireJS