updates and improvements for QLS 2025

[Remi-Gau]

Clinic

• Use the install clinic to briefly show how to create, activate, deactivate environments with conda.
• Set up SSH on github to make it easier to push: https://hyper-neuro.slack.com/archives/C06EGRL8ND7/p1715742060062759?thread_ts=1715742060.062759&cid=C06EGRL8ND7

Python

• split the module with one for beginners and one more advanced

• Things to mention as it is confusing a few people.

  • Import a whole module or just a function from that module.

ˋˋˋpython
import foo
a = 1
foo.bar(1)
ˋˋˋ

Same as
ˋˋˋpython
from foo import bar
a = 1
bar(1)
ˋˋˋ

• When writing python code for other modules, try to rely on coding syntax that students may be familiar from other languages and avoid "pythonic idioms" or "one liners". For example

  • Increment operator +=

  • List / dict comprehension

  • One liner with Tuple unpacking on returned values
    ˋˋˋpython
    import foo

    a = foo.bar(b)[0]
    ˋˋˋ

    instead of

    ˋˋˋpython
    from foo import bar

    tmp = bar(b)
    a = tmp[0]
    ˋˋˋ

  • Also where possible name all paramaters when calling a function foo(model=a, parameters=b) instead of f(a, b)

Visualization

• earlier in the week
• present higher level lib (plotly, seaborn, bokeh...)
• Several users have had issues with plotting via wsl. Getting some qt warming and no image showing.
  For some this worked
  ˋˋˋ
  conda uninstall matplotlib
  pip install matplolib
  ˋˋˋ

Misc

• reduce amount
• try to have terminal and slides on a single screen and avoid switching
• for windows users, maybe consider NOT using WSL unless for the bash section of the course: these people are likely to be using python NOT via WSL on an everyday basis, I am not sure it makes sense to add more hoops for them to jump through by forcing them to use WSL (also because WSL can cause extra problems and slow these people during the course)
• see: https://hyper-neuro.slack.com/archives/C06EGRL8ND7/p1715870869762009