---
title: "Transport routing with stplanr"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Transport routing with stplanr}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

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

```{r setup}
library(stplanr)
```

# Introduction

Routing is the process of identifying routes that enable movement between two geographic locations along the shortest path (based on mode-specific routing profiles) or in some other 'optimal' way, based on route network data.
Most open routing engines rely on OpenStreetMap (OSM) data.

We will use the example of the Isle of Wight to demonstrate routing engines.
To get OSM data for the Isle of Wight you can run the following commands:

```{r, eval=FALSE}
remotes::install_github("itsleeds/geofabrik")
library(geofabrik)
roads_iow = get_geofabrik(name = "Isle of Wight")
f = gf_filename("Isle of Wight")
file.copy(f, "iow.pbf")
options(osrm.server = "https://0.0.0.0:5000/", osrm.profile = "driving")
```


# OSRM

Routing services such as OpenStreetMap Routing Machine (OSRM) require an input network, usually from OSM.

We will use the `osrm` package:

```{r}
library(osrm)
```

In the system terminal run the following commands to make the [OSRM docker image](https://hub.docker.com/r/osrm/osrm-backend/) work for you.

```{r, engine='bash', eval=FALSE}
docker run -t -v "${PWD}:/data" osrm/osrm-backend osrm-extract -p /opt/car.lua /data/iow.pbf
docker run -t -v "${PWD}:/data" osrm/osrm-backend osrm-partition /data/iow.osrm
docker run -t -v "${PWD}:/data" osrm/osrm-backend osrm-customize /data/iow.osrm
docker run -t -i -p 5000:5000 -v "${PWD}:/data" osrm/osrm-backend osrm-routed --algorithm mld /data/iow.osrm
curl "https://127.0.0.1:5000/route/v1/driving/13.388860,52.517037;13.385983,52.496891?steps=true"
```

Now we can do routing in R!

On a single route:

```{r, eval=FALSE}
l = pct::wight_lines_30
p = line2points(l)
r = osrm::osrmRoute(src = p[1, ], dst = p[2, ], returnclass = "sf", overview = "full")
plot(r)
```

```{r, echo=FALSE}
knitr::include_graphics("https://user-images.githubusercontent.com/1825120/86902789-577d1080-c106-11ea-91df-8d0180931562.png")
```

And to find many routes via the `route()` function, resulting in something like the figure below.

```{r, eval=FALSE}
routes_osrm = route(l = l, route_fun = osrmRoute, returnclass = "sf", overview = "full")
rnet_osrm = overline(routes_osrm, attrib = "bicycle")
mapview::mapview(rnet_osrm, lwd = rnet_osrm$bicycle / 10)
```

```{r, eval=FALSE, echo=FALSE}
system.time({
  routes_osrm = route(l = l, route_fun = osrmRoute, returnclass = "sf", overview = "full")
})
30 / 0.9 # around 30 routes per second
saveRDS(routes_osrm, "routes_osrm.Rds")
piggyback::pb_upload("routes_osrm.Rds")
```


```{r, echo=FALSE}
knitr::include_graphics("https://user-images.githubusercontent.com/1825120/86858225-2970df80-c0b8-11ea-8394-07f98f1c8e8a.png")
```

```{r, eval=FALSE}
# tidy up
f = list.files(pattern = "iow")
unlink(x = f, recursive = TRUE)
```

Shut down the docker container.


```{r, eval=FALSE, engine='zsh'}
docker ps
docker stop stupefied_hopper
```