[MDBF-56] Packaging of maria Jupyter kernel (installation, artwork) Created: 2020-05-25  Updated: 2022-02-01  Due: 2020-06-12

Status: Closed
Project: MariaDB Foundation Development
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Task Priority: Major
Reporter: Kaj Arnö Assignee: Robert Bindar
Resolution: Unresolved Votes: 0
Labels: jupyter
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
PartOf
includes MDBF-62 Prepare specs for artwork Open
includes MDBF-63 Implement proper icon and packaging a... Open
is part of MDBF-53 Release first “minimum viable” Jupyte... Open

 Description   
  • packaging to be done by conda; conda packaged kernels are the easiest to use (something Robert found out when testing other kernels)
  • conda emulates a fresh Python environment; usually, most Python environments of developers are somehow broken or dirty
  • the conda installations do not go into the main Python code base of the computer
  • “you first pip install conda and then you conda install stuff”
  • we avoid a lot of complexity by not using a Python connector at all, but connecting to MariaDB through the mariadb command line
  • for this purpose, we import pexpect

 Comments   
Comment by Vicențiu Ciorbaru [ 2020-06-01 ]

Conda, unlike pip also tracks dependencies outside of python and creates a virtualenv to install packages in. Do we include mariadb-client inside this conda package? What about the MariaDB Server?

Comment by Kaj Arnö [ 2020-06-03 ]

Pro MariaDB Client

  • we know exactly which client it is
  • it isn't very costly MB-wise, anyway

Con MariaDB Client

  • slightly increased package size

Pro MariaDB Server

  • we can increase the adoption of MariaDB 10.5 this way

Con MariaDB Server

  • considerably increased package size
Comment by Robert Bindar [ 2020-06-03 ]

I need to research this, I don't have enough info to answer now. Another option is to package MariaDB for conda, it seems there is already a mysql package in conda-forge, see https://conda-forge.org/feedstocks/ and https://github.com/conda-forge/mysql-feedstock

Also we need to debate if this kind of issues should be considered for the first release or should be postponed as nice-to-have for future releases.

Comment by Kaj Arnö [ 2020-06-03 ]

I am fine with nice-to-have, for future releases!

Generated at Thu Feb 08 03:35:08 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.