ITSECL

ITSECL is a Typescript/ECL kernel for the Jupyter notebook... It provides a prototype implementation of HPCC ECL for Jupyter Notebook by referencing ITypeScript

The HPCC Systems server platform is a free, open source, massively scalable platform for big data analytics. Download the HPCC Systems server platform now and take the reins of the same core technology that LexisNexis has used for over a decade to analyze massive data sets for its customers in industry, law enforcement, government, and science.

For more information and related downloads for HPCC Systems Products, please visit https://hpccsystems.com

Installations

Prerequites: python-dev, python3.5-dev, python-pip python3-pip jupyter nodejs

For example on Ubuntu 16.04 xenial

sudo apt-get install -y curl git python-dev python3.5-dev python-pip python3-pip
sudo pip3 install jupyter
sudo curl -sL https://deb.nodesource.com/setup_8.x | sudo -E bash -
sudo apt-get install -y nodejs

On ubuntu 18.04 replace python3.5-dev with python3.6-dev

Install ITSECL

Install ITSECL globally

You don't need use "sudo" but it is usually required to configure the npm default directory

mkdir ~/.npm-global
npm config set prefix '~/.npm-global'
echo "export PATH=~/.npm-global/bin:\$PATH" >> ~/.profile
source ~/.profile

If run as root or sudo following probably is needed as root before run npm install npm config set unsafe-perm true

To install ITSECL:

npm install -g itsecl

To start Jupyter Notebook with TSECL engine:

itsecl

Install ITSECL locally

In an empty directory

npm init
npm install itsecl

To start Jupyter Notebook with TSECL engine:

./node_modules/.bin/itsecl

Alternatively way to jupyter notebook

Install itsecl kernel

itsecl --ts-install=global

# If you want to see itsecl kernel debug message add "--ts-debug"
# To verify ITSECL kernel (jsecl) installed in jupyter:
jupyter kernelspec list. "tsecl" is the kernel name.

To run ITSECL in jupyter notebook in your work directory which can be any directory (The Jupyter notebook files will be saved there):

jupyter notebook

# Or with debug information for tsecl kernel:

itsecl --ts-debug
# Then select kernel "HPCC ECL - TSECL" from "new"

To run ITSECL in jupyter console (Currenly support for console is limited. Try to use notebook instead).

jupyter console --kernel=tsecl

Usage

Run normal Javascript/Typescript code:

//JS
var i = 10; i;

Test connection with HPCC esp server

In order to run ECL code you should test connection with HPCC ESP server first

//CONN  ip=192.168.56.100; port=8010; cluster=hthor; user=<username>; password=<password>;

The above connection parameters can be provided from a file

//CONN  file=/tmp/esp.conf;

To use https set "secure=true"

Sample esp.conf:

ip=190.29.2.11
port=8018
secure=false
cluster=thor
user=hpccuser
passwd=mypassword
default=ECL

You also can add workspace directory (workspace=;) if it is not in standard space, which is the directory that Jupyter server is started. This only applys when you have HPCC Clienttools installed on the Jupyter server system.

To display the current configuration:

//CONF

Test connection with HPCC esp server

This is still under development

ECL code

//ECL
OUTPUT("Hello ECL");

If default action is set to "ECL", which is default, "//ECL" can be avoid"

To change cluster

//ECL cluster=roxie;
...


To allow the code re-usable
```sh
///<module>/<your ecl file name>

The will be saved under workspace directory. You can follow ECL syntax to includ or import your ECL code. You also can import existing code related to the workspace directory.

For example,

Cell one

///Moudle1/Def1.ecl save only
EXPORT Def1 := 10;

When click "run" it only saves the file under workspace

Cell two

IMPORT Module1;
Def2 := Module1.Def1 + 12;
OUTPUT('The value is ' + Def2)

When click "run" it should display "The value is 22".

Stop Jupyter Notebook

Clt-C and type 'y'

Build and Publish

To build

git clone https://github.com/hpcc-systems/jupyter-itsecl
cd jupyter-itsecl
npm install
npm install typescript -g
tsc

*.js files will be generated in bin and lib direcotries.

To publish to npm repository

npm login
npm publish

## ITSECL Docker Image
For using and building ITSECL Docker Image reference docker/README.md

## Troubshooting
* ERROR message with Chrome browser
ERROR:gpu_process_transport_factory.cc(1017)] Lost UI shared context.
ERROR:command_buffer_proxy_impl.cc(114)] ContextResult::kFatalFailure: Shared memory handle is not valid

To avoid it before start jupyter run:
```sh
export BROWSER=google-chrome