pprzetacznik / IElixir

Jupyter's kernel for Elixir programming language
Apache License 2.0
359 stars 42 forks source link
elixir jupyter jupyter-kernels jupyter-notebook repl

IElixir

Jupyter's kernel for Elixir

Build Status IElixir CI IElixir Docker Inline docs Coverage Status Join the chat at https://gitter.im/pprzetacznik/IElixir

Hex: https://hex.pm/packages/ielixir.

Please see generated documentation for implementation details: http://hexdocs.pm/ielixir/.

Getting Started

Table of contents

Configure Jupyter

I recommend you to use virtualenv and virtualenvwrapper for this project to isolate dependencies between this and other projects however you may also work without this if you don't like this.

$ pip install virtualenv virtualenvwrapper

Now you need to load virtualenvwrapper.sh script into your current environment. I recommend you to add this like as well to the ~/.bash_profile.sh script to have this script loaded every time you open fresh bash.

$ source /usr/local/bin/virtualenvwrapper.sh

Now using our new tools we can easily create isolated virtual environment for jupyter installation.

$ mkvirtualenv jupyter-env
$ workon jupyter-env
(jupyter-env) $ pip install jupyter

Configure IElixir

Clone IElixir repository and prepare the project

$ git clone https://github.com/pprzetacznik/IElixir.git
$ cd IElixir
$ mix deps.get
$ mix test
$ MIX_ENV=prod mix compile

If you're meeting problem with missing zeromq header files then you should install it. More.

Running all tests, including longer ones that requires more time for evaluation:

$ mix test --include skip

There may be also need to install rebar before IElixir installation, you can do this with command:

mix local.rebar --force

After this you may need to add ~/.mix/ to your $PATH variable if you don't have rebar visible yet outside ~/.mix/ directory.

Install Kernel

Simply run installation script to create file kernel.json file in ./resouces directory and bind it to the jupyter:

$ ./install_script.sh

Use IElixir

Run Jupyter console with following line:

(jupyter-env) $ jupyter console --kernel ielixir

To quit IElixir type Ctrl-D.

Run Jupyter Notebook with following line:

(jupyter-env) $ jupyter notebook resources/example.ipynb

Go to http://localhost:8888/ site (by default) in your browser and pick IElixir kernel:

Pick IElixir

Evaluate some commands in your new notebook:

IElixir basics

Magic commands

You can also use ans variable to access output of previous cell. Moreover, if you want to access any cell which you can do it using it's number by calling out map, eg. out[1].

Package management with Boyle

You can manage your packages in runtime with Boyle. Name of the package honours remarkable chemist, Robert Boyle. This package allows you to manage your Elixir virtual enviromnent without need of restarting erlang virtual machine. Boyle installs environment into ./envs/you_new_environment directory and creates new mix project there with requested dependencies. It keeps takes care of fetching, compiling and loading/unloading modules from dependencies list of that environment.

You can also use this environment as a separate mix project and run it interactively with iex -S mix from the environment directory.

Creating new Elixir virtual environment

iex> Boyle.mk("my_new_environment")
{:ok, ["my_new_environment"]}

List available virtual environments

iex> Boyle.list()
{:ok, ["my_new_environment"]}

Activate virtual environment

iex> Boyle.activate("my_new_environment")
:ok

Install new package in virtual environment and use new package

iex> Boyle.install({:number, "~> 0.5.7"})
:ok
iex> Number.Currency.number_to_currency(2034.46)
"$2,034.46"

Deactivate virtual environment and unload packages installed within that virtual environment

iex> Boyle.deactivate()
:ok
iex> Number.Currency.number_to_currency(2034.46)
** %UndefinedFunctionError{arity: 1, exports: nil, function: :number_to_currency, module: Number.Currency, reason: nil}

Additional resources:

Developement mode

If you want to see requests passing logs please use dev environment to see what is happening in the background.

(jupyter-env) $ MIX_ENV=dev jupyter console --kernel ielixir

Generate documentation

Run following command and see doc directory for generated documentation in HTML:

$ MIX_ENV=docs mix docs

Docker

You can find docker image at pprzetacznik/ielixir.

Running jupyter notebook:

$ docker run -p 8888:8888 --hostname localhost -v /my/workspace:/home/jovyan/work pprzetacznik/ielixir

Docker image is based on following images:

If you would like to make some changes to the images you can find dockerfiles in:

Elixir builds taken from hex.pm builds.txt.

Other docker images worth seeing

Some issues

ZeroMQ header files missing

===> /home/travis/build/bryanhuntesl/IElixir/deps/erlzmq/c_src/erlzmq_nif.c:24:17: fatal error: zmq.h: No such file or directory
 #include "zmq.h"
                 ^
compilation terminated.

Install ZeroMQ development package for you operating system.

Erlang configuration

There may be need to run IElixir kernel with specific erlang attribute which can be turned on by setting variable:

ELIXIR_ERL_OPTIONS="-smp enable"

This option has been included inside install_script.sh and start_script.sh scripts.

Listing all available jupyter kernels.

$ jupyter kernelspec list

Contribution

References

Some useful articles:

I was inspired by following codes and articles:

License

Copyright 2015 Piotr Przetacznik. IElixir source code is released under Apache 2 License. Check NOTICE and LICENSE files for more information.