Create model

[vladh]

TODO: Flesh this out

1. Top-10k python packages by downloads -> github repos
2. GH repos -> dep trees (pypi or meson package names)
3. Repeat (1) for pypi packages from (2)
4. Build model based on data from (1)-(3)
5. [Opt] Meson packages -> github repos

[jring-o]

Script for finding top 10k python packages by downloads:

https://github.com/jring-o/software-finder

[vinogradovkonst]

Input data for the task 1 : https://github.com/osendowment/model/blob/main/data/top-pypi-packages.csv

The dataset has all PyPI packages with at least 1M/year avg downloads in 2020-2024 ~ 5500 packages we can focus on.