Jupyter Book: Interactive books running in the cloud

I talk about the Jupyter Project, an open community that builds public tools for solving data analytics problems. I also discuss a new project, Jupyter Book, that lets you weave multiple Jupyter Notebooks together into a single book-like document.

Transcript

1. @choldgraf bit.ly/jupyter-book-strata2019
2. @choldgraf Cognitive Neuroscience Open Source
3. @choldgraf Research and Open Source Education and Open Source Jupyter @ Berkeley
6. You San Jose Coffee!
7. You San Jose Coffee! One option: walk there by myself Another option: pay somebody to drive me My favorite option: use public infrastructure
8. You San Jose Coffee!
9. BART Bus You San Jose Coffee!
10. BART Bus The “last mile” You San Jose Coffee!
13. You Your Awesome report
14. One option: build it from scratch Another option: pay somebody for a product My favorite option: build on open, modular tools You Your Awesome report
15. You Your awesome report server .ipynb
18. Here’s what this means at Berkeley...
19. • Teach students the concepts of statistical thinking • Provide them an early experience in coding and interactive computing • Teach an open stack of languages and tools • Facilitate the learning experience with shared infrastructure • Work in partnership with open communities like Jupyter, contribute our improvements upstream.
21. Us Interactive course content for Data Science
22. Us Content structure Analytics Tools User Interface Coding Environments Interactive Data Science course content
23. Us .ipynb Interactive Data Science course content
24. • A way to display notebook content in a more familiar, beautiful, and accessible form. • A way to add more bells and whistles that are geared towards reading. • An ability to connect this content with interactive computation. • A way to deploy all of this quickly and easily.
25. Us .ipynb Interactive Data Science course content
26. jupyter.org/jupyter-book
28. 1. A collection of Jupyter Notebooks + markdown files 2. A Table of Contents yaml file (toc.yaml) 3. Some optional configuration (config.yaml) jupyter-book create mybook --content-folder path/to/content \ --config myconfig.yaml \ --toc mytoc.yaml
29. mybook/ ├── assets ├── _config.yml ├── content ├── _data ├── Gemfile ├── Guardfile ├── _includes ├── _layouts ├── Makefile ├── requirements.txt ├── _sass ├── scripts └── _site
30. • Build a static website out of a single set of text files • Use a templating engine to interpolate content into a site structure. • Lots and lots (and lots) of options to choose from.
31. • Build a static website out of a single set of text files • Use a templating engine to interpolate content into a site structure. • Lots and lots (and lots) of options to choose from. ✨ ✨ ✨ ✨ ✨ ✨ ✨ ✨ ✨
32. • Make sure we retain important metadata about cells • Need to keep structure for cells, inputs, outputs • Want rich content! jupyter-book build mybook
33. mybook/ ├── assets └──_build ├── features │ ├── features.md │ ├── interact.md │ ├── markdown.md │ ├── notebooks.md │ └── search.md ├── guide │ ├── 01_overview.md │ ├── 02_create.md └── intro.md jupyter-book build mybook
34. • You can either upload to GitHub and build it with GitHub Pages • Or, install Ruby locally and build the HTML yourself! cd mybook make serve
35. mypage.ipynb mypage.md mypage.html
37. {"metadata" : { "kernel_info": { "name" : "my awesome kernel" }, "language_info": { "name" : "python3", "version": "0.1"} }, "nbformat": 4, "nbformat_minor": 0, "cells" : [ <list-of-cell-dictionaries> ]} Notebook metadata [ {'cell_type': 'markdown', 'metadata': {}, 'source': '# Hi there'}, {'cell_type': 'code', 'execution_count': 10, 'metadata': {}, 'outputs': [{'name': 'stdout', 'output_type': 'stream', 'text': '[-0.61023336 -0.35754269 -0.18629438 0.84355958 -0.14921895]\n'}], 'source': 'import numpy as np\ndata = np.random.randn(5)\nprint(data)'}, {'cell_type': 'markdown', 'metadata': {}, 'source': 'Look at those numbers!'} ] Cell Data
38. https://jupyter.org/jupyter-book/features/hiding.html
39. Connecting with the cloud
40. jupyter.org/hub
41. mybinder.org
42. • A page built from a notebook gets an interact button • Use JupyterHub/Binder to spawn a Jupyter server • `git pull` the underlying notebook for the page • Initialize an interactive environment
43. Chris Is Trying A Live Demo Hopefully he doesn’t embarrass himself too badly. https://www.inferentialthinking.com/chapters/13/3/Confidence_Intervals.html
46. • JupyterHub and Binder both run Jupyter kernels in the cloud • We can connect to these kernels on the backend turn “static” webpages into dynamic pages.
47. thebelab ◦ Converts code blocks into editable cells that connect to a Binder / JupyterHub kernel ◦ http://thebelab.readthedocs.org nbinteract ◦ Display widgets in a static webpage, backed by a Binder kernel ◦ https://www.nbinteract.com/ ✨experimental✨
48. Chris Is Trying A Live Demo Hopefully he doesn’t embarrass himself too badly. https://jupyter.org/jupyter-book/features/interactive_cells.html
50. https://github.com/simpeg-research/heagy-2018-aem
51. • Better story around citations, multiple output types, etc. pandoc mynotebook.ipynb -o mypage.html --filter pandoc-citeproc --bibliography=mybib.bibtex
53. • I am not a web developer!
54. Launches in a month
55. jupyter.org/jupyter-book Us .ipynb Interactive Data Science course content
56. jupyter.org/jupyter-book Us .ipynb Interactive Data Science course content
57. @choldgraf github.com/jupyter/jupyter-book discourse.jupyter.org • • •
58. @choldgraf http://bit.ly/jupyter-book-strata2019