How long is the Wildland–Urban Interface boundary? At first glance, this sounds like a straightforward measurement problem: draw the line separating developed places from wildland vegetation and compute its length. But that apparent simplicity collapses as soon as method enters the picture. The measured boundary depends on what we decide counts as the boundary in the first place, and on the scale at which we measure it.
This repository is organized around that scientific problem. It treats WUI boundary length not as a single fixed quantity to be reported once, but as a value conditioned by assumptions and resolution. The project combines narrative scientific interpretation, interactive web-based exploration, and a compact analytics workflow so that the conceptual argument and the computational logic remain tied together.
The intuition is familiar from the coastline paradox: as measurement becomes finer, irregular boundaries often appear longer. The WUI boundary can behave similarly. A coarser ruler smooths local complexity and yields a shorter estimate; a finer ruler resolves more structure and yields a longer one.
A second source of variation is delineation. Changing the definition of what counts as "WUI boundary" changes the object being measured, not just the instrument used to measure it. In other words, there are two different levers: one changes the boundary itself, and the other changes the scale of measurement applied to that boundary.
In this project, measured boundary length is written as (L_d(\epsilon)), where (d) denotes delineation choice (the boundary definition) and (\epsilon) denotes measurement scale (effective ruler length or resolution). The central claim is that WUI boundary length should be treated as a scale-conditioned quantity, not a fixed scalar.
This repository contains a scientific website in docs/ that presents the argument as a narrative investigation rather than as a static reference page. It also includes interactive prototypes in docs/ui-drafts/ that help communicate scale dependence and definition dependence in a more exploratory format. Alongside those materials is a small analytics framework in src/boundary_analytics/ and executable scripts in scripts/ that demonstrate the scaling logic with synthetic geometries.
The current code-and-docs pairing is intentionally explicit: the website explains the reasoning, and the scripts show how (L_d(\epsilon)) can be produced in practice. Outputs from demo runs are written to outputs/ and selected artifacts are mirrored into docs/assets/ for web presentation.
For extended narrative context (motivation, hypothesis development, and core dataset framing), see PROJECT_CONTEXT.md.
The core layout is compact. docs/ contains the website narrative, while docs/ui-drafts/ holds interactive draft experiments. src/wui_boundary/ is not used in this repository; the analytics code currently lives in src/boundary_analytics/. scripts/ contains runnable analysis entry points. outputs/ stores generated demo artifacts. tests/ includes lightweight checks for the analytics scaffold. Site configuration lives in mkdocs.yml.
python -m pip install -e .
python scripts/run_minimal_demo.pyDemo outputs are written to outputs/ (specifically outputs/minimal_demo/ for the current minimal run). The canonical runtime install path is editable package install (python -m pip install -e .), which includes geospatial dependencies for the streaming real-data pilot. requirements.txt is reserved for documentation tooling (MkDocs) layered on top when needed.
The repository includes a lightweight Playwright reviewer that checks the deployed GitHub Pages site at:
https://cu-esiil.github.io/WUI_boundary/
It runs on pull requests and can also be launched manually from the Actions tab using the Playwright Site Review workflow. The review focuses on key scientific pages and catches common publication failures (broken page loads, missing figure/table/iframe content, console errors, and obvious navigation regressions).
Artifacts from each run are available in GitHub Actions:
playwright-report(HTML report)playwright-test-results(screenshots, traces on failure, and JSON diagnostics for console/request issues)
You can also run it locally:
npm install
npx playwright install --with-deps chromium
npm run test:site-reviewRun this before opening or proposing a PR for website-related changes:
npm run review:siteThis one command runs the site-facing analysis step (python scripts/run_minimal_demo.py), performs mkdocs build --strict, ensures Playwright dependencies are installed, and executes local Playwright tests.
If Playwright fails, inspect artifacts and fix before retrying:
playwright-report/for the HTML reporttest-results/for screenshots, traces, and videos on failure
GitHub Actions remains the final CI gate, but local review should catch most website/UI regressions earlier.
At present, the repository demonstrates the conceptual framework with synthetic boundary geometries and synthetic scaling outputs. It is designed to grow toward empirical WUI datasets and case-based analyses while keeping delineation choices, measurement scale, and interpretation transparent.