Paper and experiment workspace for PolicyEngine's L0 dataset-reduction work on the Populace stack.
The initial manuscript narrative was built in relation to the old
PolicyEngine/policyengine-us-data paper draft. It remains LaTeX-first for now,
with a Quarto entry point that includes the existing section files so the paper
can be rendered and revised incrementally while the implementation moves to
PolicyEngine/populace.
paper/index.qmd- Quarto entry point for the paper.paper/main.tex- legacy LaTeX root copied from the prior draft.paper/sections/- manuscript sections.paper/tables/- generated/result tables included by the paper.paper/figures/- figure outputs.paper/bibliography/references.bib- references.src/l0_paper/- Python helpers for Populace-backed experiments.tests/- tests that exercise the active Populace API surface.experiments/- notebooks/configs/result scripts for paper experiments.
This repo expects to be cloned next to PolicyEngine/populace:
PolicyEngine/
l0-paper/
populace/
The pyproject.toml points uv at the sibling Populace packages, so local
tests use the active checkout rather than archived Microplex internals.
uv sync --all-extras --dev
uv run pytest
uv run ruff check .The smoke test runs a tiny populace.calibrate.calibrate(..., target_records=...) L0 calibration to verify the experiment repo is wired to
Populace's implementation.
Install Quarto, then run:
quarto renderThe current local machine does not have Quarto installed. The legacy LaTeX source can still be checked with:
cd paper
latexmk -pdf main.texNew code in this repository should target active Populace APIs rather than
archived microplex or microplex-us internals. The imported manuscript still
contains old Microplex wording in places; updating the narrative to Populace is
part of the next migration pass.