---
title: "Comparing evacuation scenarios with evacpath"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Comparing evacuation scenarios with evacpath}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(collapse = TRUE, comment = "#>", eval = FALSE)
```

`evacpath` can compare pedestrian-evacuation assumptions without changing the
underlying study inputs. This article uses the packaged example data and clips
all spatial inputs to a small focus area before modeling.

## Load example inputs

```{r load-inputs}
library(evacpath)
library(terra)

focus <- ext(669000, 675000, 9223000, 9225000)

dem <- crop(rast(system.file("extdata/dem.tif", package = "evacpath")), focus)
inundation <- crop(
  rast(system.file("extdata/tsunami_inundation_depth.tif", package = "evacpath")),
  focus
)
roads <- crop(vect(system.file("extdata/rds.gpkg", package = "evacpath")), focus)

zones <- prepare_tsunami_zones(
  inundation = inundation,
  dem = dem,
  target_crs = "EPSG:32748"
)

roads <- clean_roads(
  roads,
  exclude = list(field = "man_made", values = "pier"),
  target_crs = "EPSG:32748"
)
```

## Run a baseline model

`walking_speed_mps` controls conversion from route distance to evacuation time.
The `lcp_*` arguments control the conductance surface and can change modeled
route geometry.

```{r baseline}
baseline <- run_evacpath(
  hazard_zone = zones$hazard_zone,
  escape_zone = zones$escape_zone,
  roads = roads,
  dem = zones$dem,
  target_crs = "EPSG:32748",
  max_origins = 500,
  max_destinations = 100,
  walking_speed_mps = 1.22,
  keep_routes = TRUE
)
```

## Compare walking speed and path assumptions

```{r compare}
comparison <- compare_evac_scenarios(
  hazard_zone = zones$hazard_zone,
  escape_zone = zones$escape_zone,
  roads = roads,
  dem = zones$dem,
  target_crs = "EPSG:32748",
  max_origins = 500,
  max_destinations = 100,
  scenarios = list(
    baseline = list(walking_speed_mps = 1.22),
    slow_walkers = list(walking_speed_mps = 0.75),
    alternative_neighbours = list(lcp_neighbours = 8)
  )
)

comparison$summary
```

## Diagnose inputs

Diagnostics collect coordinate reference system, geometry, overlap, elevation,
origin, destination, and optional reachability issues in one report.

```{r diagnose}
diagnostics <- diagnose_evac_model(
  hazard_zone = baseline$hazard_zone,
  roads = baseline$roads,
  dem = baseline$dem,
  escape_zone = baseline$escape_zone,
  origins = baseline$road_points,
  destinations = baseline$escape_points,
  conductance = baseline$conductance
)

diagnostics
has_errors(diagnostics)
```

## Map bottlenecks

Routes are retained only when `keep_routes = TRUE`. This keeps ordinary runs
lightweight while allowing route-density analysis when it is needed.

```{r bottlenecks}
bottlenecks <- map_evac_bottlenecks(evac_result = baseline)

plot(bottlenecks$density_raster)
plot(bottlenecks$high_density_polygons, add = TRUE)
```