Blog · almond

almond 0.3.1 was just released. This release contains updates along with a fix to have almond work on Colaboratory, thanks to the only Shadaj Laddad!

Updates

This release switches to Ammonite 1.6.4, fs2 1.0.4, and jeromq 0.5.0, among others.

Ammonite 1.6.4 mainly replaced scalaj-http by requests-scala in the classpath it exposes. It also doesn't print back private values, e.g.

private val n = 2
val m = n + 1

just prints back

m: Int = 3

rather than

n: Int = 2
m: Int = 3

(n isn't accessible in subsequent code, so it doesn't make much sense to pretty-print it.)

Colaboratory

Shadaj Laddad contributed a fix allowing almond to run fine on Colaboratory.

For the braves, the following notebook (more stable link) allows one to setup almond in a Colaboratory environment.

As using custom kernels isn't an official feature of Colaboratory yet, one cannot start almond sessions straightaway. Instead, an almond notebook has to be imported. Imported notebooks will have the right kernel. For example, notebooks from the almond examples repository, can be imported, and modified / run, once almond is installed.

almond 0.3.0 was just released. This release contains mostly updates, but also notable changes in the build of almond, that can be of interest to you!

Updates

This release switches to Ammonite 1.6.3, and ammonite-spark 0.3.0. Apart from a few internal changes, these were mostly maintaince releases.

There were some minor binary incompatible changes in those, justifying the minor version bump here.

One-liner to run from sources

The previous "run from sources" instructions relied on publishing almond locally, then generating a coursier boostrap to run the locally published artifacts. This was a bit cumbersome overall, requiring running commands both inside sbt and outside of it.

The almond build now defines sbt tasks making it more practical. These allow to run almond from sources in just one command, like

$ sbt jupyterStart shell

See the install from sources page for more details.

New front page

The almond website now has a shiny new front page. It draws inspiration from the numerous docusaurus users out there, most notably the scalameta projects. Thanks to them!

almond-spark for scala 2.12

The almond-spark module is now published for scala 2.12, so that one can rely on it from an almond kernel running scala 2.12, when using spark 2.4.0.

Docker images

Docker images for almond, based on official Jupyter images, are now automatically generated and pushed upon each Almond release. These are based on previous work by @Atry, and live in this repository.

For example you can run almond 0.3.0 with

$ docker run -it --rm -p 8888:8888 almondsh/almond:0.3.0

Then copy the token from the Jupyter logs, and head to http://localhost:8888.

See other the images and instructions here.