# 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(RCzechia)                         # For Czech administrative regions data
library(ggplot2)                          # For data visualization
library(ggrepel)                          # For text labeling in plots

# 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 the location "Vráž" to "Vráž (okres Písek)"
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)

# Convert latitude and longitude coordinates to spatial features
poi <- lat_longs %>%
  st_as_sf(coords = c("longitude", "latitude"), crs = 4156)

# Create a base plot of the Czech Republic
plot1 <- ggplot() +
  geom_sf(data = republika("low"), color = "gray30", fill = NA) + # Outline of Czech Republic
  geom_sf(
    data = subset(RCzechia::reky(), Major == TRUE), # Major rivers
    color = "steelblue", alpha = .7
  ) +
  # Plot spa towns on the map
  geom_sf(data = poi, color = "red", shape = 1, size = 1) +
  # Label spa towns on the map
  geom_text_repel(data = poi, aes(lat_longs$longitude, lat_longs$latitude, label = Lokalita),
                  color = "black", size = 2.5) +
  labs(title = "Lázeňská místa v ČR") +    # Set title of the plot
  theme_minimal() +                         # Set minimal theme for the plot
  theme(axis.title = element_blank())       # Remove axis titles

# Save the plot as an image file
ggsave("lazne.jpg", bg = "white", width = 30, units = "cm")