A more performant notebook? #331
Description
The quarto notebook at the moment uses MultiQC as a library which automatically parses logs, which in turn loads a lot of other libraries which might cause issues, in particular on arm chips.
It's likely though that we could make a more performant notebook by not using MultiQC. Data could be plotted directly and perhaps dynamically, and using more performant libraries, like polars, etc.
Alternatively/Additionally, we could make it a pure Marimo notebook, since we're currently purely using Python. This would give us access to lots of extra features, skip using Jupyter under the hood ( which is part of the issue in #325 ), and allow us to supply a notebook to the user to further explore results (which would also entail packaging the logs somehow too).