M1 Artificial Intelligence · semester 7 · Python
Python: tooling and data
Environments and packaging first, then images, a command-line tool, and logging.
The practicals
TP1
Environments and packages
Virtual environments with venv and pip, then uv, then notebooks running against each. Built around a weather API.
TP2
Images as Numpy arrays
Loading, greyscale, padding, convolution, blurring and edge detection.
TP3
A duplicate finder on the command line
A click tool: walking a directory, hashing files, handling duplicates and corrupt images, resizing, reporting.
TP4
Logging, then a logger of your own
The standard module, then a logger class of your own: levels, file output, decorators, a static class, a singleton.
The environment
All four run in one conda environment, declared once in
environment.yml and registered as the Jupyter kernel
Python 3.11 (m1ai-programming). TP1 is the exception by design: its
subject is environments, so it builds its own with venv and
uv rather than using this one.