#!/bin/sh

# Define input and output files
url="https://unicode.org/Public/emoji/latest/emoji-test.txt"
input_file="${XDG_DOTFILES_DIR:-${HOME}/.dotfiles}/global/.local/share/thesiah/chars/emoji_raw"
temp_file="${XDG_DOTFILES_DIR:-${HOME}/.dotfiles}/global/.local/share/thesiah/chars/emoji_temp"
output_file="${XDG_DOTFILES_DIR:-${HOME}/.dotfiles}/global/.local/share/thesiah/chars/emoji"

# Create the directory for output files if it doesn't exist
mkdir -p "$(dirname "$input_file")"

# Download the emoji file
echo "Downloading emoji-test.txt from Unicode..."
if curl -o "$input_file" -L "$url"; then
  echo "Download complete! File saved to: $input_file"
else
  echo "Failed to download emoji"
  exit 1
fi

# VS16 (U+FE0F)
vs16="$(printf '\xEF\xB8\x8F')"

awk '
    /^[[:space:]]*$/ || /^#/ { next }
    !/(fully-qualified|component)/ { next }
    /200D/ { next }
    $2 ~ /1F3F[BCDEF]/ { next }
    { print }
' "$input_file" >"$temp_file"

# Extract + VS16 제거
awk -F'#' '
    {
        if (NF >= 2) {
            full_data = $2
            gsub(/^[[:space:]]+|[[:space:]]+$/, "", full_data)

            split(full_data, parts, " ")
            emoji = parts[1]

            description = ""
            for (i = 3; i <= length(parts); i++) {
                description = description parts[i] " "
            }

            gsub(/[[:space:]]+/, " ", description)
            gsub(/^[[:space:]]+|[[:space:]]+$/, "", description)

            print emoji, description
        }
    }
' "$temp_file" |
  sed "s/$vs16//g" >"$output_file"

rm -rf "$input_file" "$temp_file"
echo "Processing complete! File saved to: $output_file"