Pharo Smaltalk kernel for Jupyter. This project is implemented on Pharo 9 64 bits and test it on Mac Os X. It uses ZeroMQ ported from zeromq project to uFFI. Roassal 2 integration is supported. Main branch in this repository is in active development.
There you are a few examples on using iPharo, you can test some of them in Binder.
- Tutorial 1 - basic statistics
- Tutorial 2 - tensorflow
- Tutorial 3 - Linear regression with tensorflow and polymath
First of all install Jupyter notebooks, I suggest you to install Anaconda package. Once you have Jupyter Notebooks installed, load iPharo in a fresh Pharo image.
Metacello new
baseline: 'IPharo';
repository: 'github://jmari/IPharo:master/repository';
load:'default'
Kernel.json file should be created by Metacello (in Mac or Linux) in the correct place. If you are not able to start a new notebook in Pharo Smalltalk language, create this file manually. Create the folder '/usr/local/share/jupyter/kernels/pharo'. Create file 'kernel.json' with contents
'{
"argv": [
"/Path/To/Your/vm/Pharo",
"/Path/to/your/image/Pharo6.1-64.image",
"ipharo",
"{connection_file}"
],
"display_name": "Pharo Smalltalk",
"language": "smalltalk"
}'
Optional, copy an icon with file name logo-64x64.png.
If you are only interested in ZeroMQ Binding, please do:
Metacello new
baseline: 'IPharo';
repository: 'github://jmari/IPharo:master/repository';
load:'zmq'