Problem reading file with some leaflet content #455
Description
This comes from investigating
The problem seems to be that xml2 doesn't parse after the leaflet script part.
Here is a reprex without bookdown*
## load libraries
library(leaflet)
## open services
boem_open = arcgislayers::arc_open('https://services7.arcgis.com/G5Ma95RzqJRPKsWL/arcgis/rest/services/Wind_Lease_Boundaries__BOEM_/FeatureServer/8')
## active lease areas
active_leases = arcgislayers::arc_select(boem_open) |>
dplyr::filter(substr(LEASE_NUMBER, 1, 5) == 'OCS-A') |>
sf::st_transform(4326) |>
sf::st_make_valid()
map = leaflet() |>
addProviderTiles(providers$Esri.WorldTopoMap) |>
addPolygons(data = active_leases, fillColor = 'transparent', color = 'red', opacity = 1, weight = 2) |>
addPolygons(data = active_leases, fillColor = 'transparent', color = 'red', opacity = 1, weight = 2)
path <- tempfile(pattern = "bug", fileext = ".html")
htmlwidgets::saveWidget(map, path)
#> Registered S3 method overwritten by 'jsonlite':
#> method from
#> print.json jsonify
library(xml2)
html <- read_html(path)
# 3 last lines of the file
res <- readr::read_lines_raw(path)
vapply(res[tail(seq_along(res), 3)], rawToChar, "html content")
#> [1] "<script type=\"application/htmlwidget-sizing\" data-for=\"htmlwidget-dc2ea44e59569184d7e4\">{\"viewer\":{\"width\":\"100%\",\"height\":400,\"padding\":0,\"fill\":true},\"browser\":{\"width\":\"100%\",\"height\":400,\"padding\":0,\"fill\":true}}</script>"
#> [2] "</body>"
#> [3] "</html>"
# they can't be seen by xml2
xml2::xml_find_first(html, "//body") |> xml2::html_structure()
#> <body [style]>
#> {text}
#> <div#htmlwidget_container>
#> {text}
#> <div#htmlwidget-dc2ea44e59569184d7e4 .leaflet.html-widget.html-fill-item [style]>
#> {text}
#> {text}
#> <script [type, data-for]>
#> {cdata}
# script tag after the leaflet is not found
xml2::xml_find_all(html, "//body//script")
#> {xml_nodeset (1)}
#> [1] <script type="application/json" data-for="htmlwidget-dc2ea44e59569184d7e4 ...Created on 2025-03-14 with reprex v2.1.1
The script type="application/htmlwidget-sizing" element is not found.
This is the issue with bookdown processing as xml2 is used for post processing the html after pandoc.