Skip to content

Latest commit

 

History

History

jupyter

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
Dockerfile
Dockerfile
 
 
MLContentsManager.py
MLContentsManager.py
 
 
Readme.md
Readme.md
 
 
jupyter_notebook_config.py
jupyter_notebook_config.py
 
 

Readme.md

Introduction

The Jupyter Notebook is an open-source web application that allows you to create and share documents that contain live code, equations, visualizations and narrative text.

This extension allows you to browse and edit InterSystems IRIS BPL processes as jupyter notebooks.

Note that currently default Python 3 executor is used.

This extension assumes that annotations contain Python code and uses activities names as preceding read-only headings.

This is a direct Jupyter<->BPL bridge, at no time file representations of processes or notebooks are created.

Screenshots

Process Explorer

Process Editor

Installation

Prerequisites

  1. You'll need InterSystems IRIS 2019.2+.
  2. Install PythonGateway v0.8+ (only isc.py.util.Jupyter and isc.py.ens.ProcessUtils are required).
  3. Update ObjectScript code from the repo edge.

Automatic installation

  1. Run do ##class(isc.py.util.Jupyter).Install() and follow the prompt.

Manual installation

Manual installation contains all the same steps as automatic installation, but you need to execute them manually.

  1. Install IRISNative for Python 3.6.7 (cp3 OR py3 should be in a finename, wheel is in \dev\python\ folder inside InterSystems IRIS installation path):
pip install <IRIS>\dev\python\irisnative-*cp3*.whl
pip install <IRIS>\dev\python\*py3*.whl
  1. Install Jupyter.
pip install jupyter
  1. Check jupyter_notebook_config.py. It assumes the following defaults for IRIS connection:
  • host: localhost
  • port: 51773
  • namespace: USER
  • user: _SYSTEM
  • password: SYS
  1. If you need other connection parameters values, modify jupyter_notebook_config.py. For example to connect to InterSystems IRIS instance on port 51776 you'll need to add this line to the bottom of jupyter_notebook_config.py:
c.MLContentsManager.port = 51776

You can configure all settings (host, port, namespace, user, password) this way. Completely customized config looks like this:

exec(open("MLContentsManager.py").read())
c.NotebookApp.contents_manager_class = MLContentsManager
c.MLContentsManager.host = 'localhost'
c.MLContentsManager.port = 51773
c.MLContentsManager.namespace = 'USER'
c.MLContentsManager.user = '_SYSTEM'
c.MLContentsManager.password = 'SYS'

Docker

You can also run PythonGateway with Jupyter in Docker. To run the latest image execute:

docker run -d \
  -p 14558:52773 \
  -p 8888:8888 \
  --name irisj \
  intersystemsdc/irispyj:latest \
  --log $ISC_PACKAGE_INSTALLDIR/mgr/messages.log \
  --after /start.sh

You might be required to start jupyter manually, to do that execute: docker exec -d irisj sh /start.sh.

To build Docker image execute: docker build --force-rm --tag intersystemsdc/irispyj:latest .

Runnning

After completing automatic or manual installation open OS bash in the folder with jupyter_notebook_config.py and MLContentsManager.py and start jupyter with:

jupyter notebook

Notes

  • Ignore Checkpoint failed warning on save