# Load necessary libraries
library(dplyr, warn.conflicts = FALSE)     # For data manipulation
library(rvest)                             # For web scraping
library(janitor)                           # For cleaning data
library(tidygeocoder)                      # For geocoding addresses
library(sf)                                # For spatial data manipulation
library(httr)                              # For HTTP requests
library(jsonlite)                          # For JSON manipulation
library(leaflet)                           # For interactive maps
library(htmlwidgets)                       # For saving the HTML file

# Get the directory path of the currently executing script file
this_dir <- dirname(parent.frame(2)$ofile)

# Set the working directory to the directory path obtained above
setwd(this_dir)

# Define the URL of the webpage containing information about spa towns in the Czech Republic
url <- "https://www.mzcr.cz/seznam-lazenskych-mist-v-cr/"

# Read the HTML content of the webpage
html <- read_html(url)

# Extract the table from the HTML content
table <- html |> 
  html_element("table") |> 
  html_table()

# Use the first row of the table as column names
table <- table %>%
  row_to_names(1)

# Clean up the data: Rename certain locations for consistency
table <- table %>%
  mutate(Lokalita = case_when(
    Lokalita == "Vráž" ~ "Vráž (okres Písek)",
    TRUE ~ Lokalita
  ))

# Geocode the locations in the table to obtain latitude and longitude coordinates
lat_longs <- table %>%
  tidygeocoder::geocode(Lokalita, method = "arcgis", lat = latitude, long = longitude)

# Further clean up the data: Adjust certain location names for consistency
lat_longs <- lat_longs %>%
  mutate(Lokalita = case_when(
    Lokalita == "Bílina" ~ "Bílina (město)",
    Lokalita == "Bludov" ~ "Bludov (okres_Šumperk)",
    Lokalita == "Hodonín – Josefov" ~ "Josefov (okres Hodonín)",
    Lokalita == "Karviná – Darkov" ~ "Lázně Darkov",
    Lokalita == "Kostelec u Zlína" ~ "Kostelec (Zlín)",
    Lokalita == "Lednice" ~ "Lednice (okres Břeclav)",
    Lokalita == "Mšené Lázně" ~ "Mšené-lázně",
    Lokalita == "Teplice v Čechách" ~ "Teplice",
    Lokalita == "Toušeň" ~ "Lázně Toušeň",
    TRUE ~ Lokalita
  ))

# Create Wikipedia and API links for each location
lat_longs$wiki <- paste0("https://cs.wikipedia.org/wiki/", lat_longs$Lokalita)
lat_longs$source <- paste0("https://cs.wikipedia.org/w/api.php?action=query&titles=", lat_longs$Lokalita, "&prop=pageimages&format=json&pithumbsize=300")

# Function to get thumbnail link for a place from Wikipedia API
get_thumbnail_link <- function(place_name) {
  api_url <- "https://cs.wikipedia.org/w/api.php"
  params <- list(
    action = "query",
    titles = place_name,
    prop = "pageimages",
    format = "json",
    pithumbsize = 300
  )
  
  response <- GET(api_url, query = params)
  content <- content(response, as = "text", encoding = "UTF-8")
  data <- fromJSON(content)
  
  if ("pages" %in% names(data$query) && length(data$query$pages) > 0) {
    page_id <- names(data$query$pages)[1]
    thumbnail_link <- data$query$pages[[page_id]]$thumbnail$source
    if (!is.null(thumbnail_link)) {
      return(thumbnail_link)
    }
  }
  
  # Return NA if thumbnail link is not found
  return(NA)
}

# Add a column with thumbnail links for each location
lat_longs$thumbnail_link <- sapply(lat_longs$Lokalita, get_thumbnail_link)

# Define content for popups in leaflet map
content <- ~paste0(
  '<div style="width: 320px;">',
  '<b>', Lokalita, '</b><br/>',
  '<img src="', thumbnail_link, '"></img>',
  '<a href="', wiki, '">Wikipedie</a>',
  '</div>'
)

# Create an interactive leaflet map with spa locations
m <- leaflet(data = lat_longs) %>%
  addTiles(group = "OSM") %>%
  addProviderTiles(providers$OpenTopoMap, group = "Topo") %>%
  addLayersControl(baseGroups = c("OSM", "Topo")) %>%
  addAwesomeMarkers(~longitude, ~latitude, popup = content, label = ~as.character(lat_longs$Lokalita))

  m

saveWidget(m, file = "lazne.html")