Awesome
PYNT (PYthon iNTeractive)
Get your code into a jupyter notebook. Anytime. Anywhere.
Check out my SciPy 2018 talk for the motivation and a feature tour of pynt.
Features
Generate jupyter notebooks on the fly
No more copy and pasting code into jupyter notebooks. Each line of code gets its own cell.
Attach a jupyter notebook to a running process
Run a command which hits the code in the notebook. Restart the notebook kernel to attach to that process.
Syntax transformations
Unroll the first pass of loops for increased interactivity.
Scroll the resulting jupyter notebook with the code buffer
Never forget which cell a code line corresponds to.
Quick Start
Disclaimer: pynt is in beta. Make sure to back-up your code before using it!
Install the codebook module with pip.
$ pip install git+https://github.com/ebanner/pynt # latest version of codebook in github
Then install pynt in emacs through MELPA.
M-x package-install RET pynt
The next time you visit a python file pynt mode will be active.
What is pynt?
pynt is an emacs minor mode for getting source code into jupyter notebooks so you can hack on it there. If you have access to source code and a command to call it with then you can get your code into a jupyter notebook.
However, just pasting your code into one big jupyter notebook cell is not particularly useful. pynt also
- splits up code into cells so it's easy to evaluate small bits
- sets up the state required to run code (by allowing you to attach notebooks to external processes)
- takes code previously buried in various namespaces (e.g. functions and loops) and exposes them to the global namespace so you can interact with them
Using pynt
It is highly recommended that you familiarize yourself with Emacs IPython Notebook (EIN) first as pynt, at its core, is a tool to make working with EIN easier.
Once you have opened a python file and pynt mode is active, cursor over to the region of code you would like to dump into a notebook and hit C-c C-s. If you need to "re-dump" the code into the notebook then hit C-c C-e.
If you want to attach a jupyter notebook to a running process, then run a command which hits the jupyter notebook code. Restart the jupyter notebook kernel with C-c C-r (ein:notebook-restart-kernel-command). When you see the message ein: [info] Starting channels WS: ... your notebook is attached!
How pynt works
pynt uses a custom kernel manager for attaching to jupyter notebook kernels started via third-party processes. When pynt generates a jupyter notebook from a code region that code region is replaced with a IPython kernel breakpoint so that subsequent commands that hit it will start a jupyter kernel for the notebook to attach to. See here for more information.
pynt also makes heavy use of the ast module to parse your code into chunks which are then dumped into notebook cells.
Related Projects
pynt is a tool that truly stands on the shoulders of giants. Here are some projects where if they had not existed, then pynt would not have been possible.