Some diagrams relating to how Jupyter works, maybe?!
Diagrams are created using a variety of techniques:
- using image generators from blockdiag.com/en/;
- using the Draw.io editor (available via JupyterLab).
This repository also includes Jupytext, which allows notebooks to be created / edited as markdown files.
The diagrams are still very sketchy, and may not be true... What I'm trying to work towards is a set of simple diagrams that are quite easy to read, and give an overview of how the various pieces firt together, and maybe some of the process steps taken by, a varety of Jupyter setups. They are not necessarily intended as documentation... If they're simplistic but not wrong, that's fine. If they are wrong, please let me know where...
To create a new notebook, create a new text file in the markdown
folder, with an .md
suffix. Write your markdown with code contained in identified code blocks. For example:
# Header
Some *markdown* text.
```python
#Some code...
print('hello')`
```
Did that work?
Save and close the file, then click on it from the Jupyter file listing to launch it, under Jupytext, as a notebook.
A dual notebook is also saved in the hidden .notebooks
directory.
To open the draw.io application, open JupyterLab (replace the \tree
path when launched from MyBinder, and click on the Diagram
button in the Launcher.
Alternatively, from the JupyterLab file browser, in the drawio
directory, double click on a .dio
file to open it in draw.io within JupyterLab.
The drawio CLI also works with folders and the
-f
flag can be used to specify the format. To convert all the files in the files/ folder to png:$ mkdir out $ draw.io -x -f png -o out/ files/
See Draw.io from the command line (In particular the "Convert multiple files" section at the end of the page)