ddg.sh - Search the internet with DuckDuckGo from your terminal

ddg.sh

#!/bin/sh
# ddg.sh - Search the internet with DuckDuckGo from your terminal
# Made with boredom, on Christmas Eve, by mftrhu
# Requirements:
#   awk, curl OR wget, sed, head, tail, grep, tput (MAYBE)

available () {
  command -v "$1" >/dev/null 2>&1
}

split () {
  awk -F "$1" "{print \$$2}" -
}

between () {
  split "$1" 2 | split "$2" 1
}

bold () {
  sed -r -e "s|<b>|$(tput bold)|g" \
         -e "s|</b>|$(tput sgr0)|g"
}

urldecode () {
  awk -niord '{ printf RT?$0chr("0x"substr(RT, 2)):$0 }' RS=%..
}

urlencode () {
  #TODO: actually do more shit here
  sed -e 's/ /\+/g'
}

htmlstrip () {
  sed -r -e "s|</?[a-zA-Z]+>||g"
}

htmldecode () {
  sed -e "s/&quot;/'/g" -e "s/&#x27;/'/g" -e 's/&amp;/\&/g' # | \
  #  awk -niord '{ printf RT?$0chr("0x"substr(RT, 4)):$0 }' RS='&#x..;'
  #BUG: awk gives up the ghost, sometimes, first seen with `-p 4 bah`
}

usage () {
  printf "Usage: %s [-hdq] [-g N] [-p P] QUERY...\n" "$0"
  printf "Performs a DuckDuckGo search and prints the results.\n" | fmt -w "$cols"
  printf "\n"
  printf "DuckDuckGo bangs are (mostly) unsupported, and will just be opened in the browser.\n" | fmt -w "$cols"
  printf "\n"
  printf "  -h    shows this help and quits\n"
  printf "  -l    opens the first result in the browser\n"
  printf "  -g N  opens the Nth result in the browser\n"
  printf "  -p P  shows the Pth page of results\n"
  printf "  -q    does not print the descriptions\n"
  printf "  -d    dumps the results as tab-separated values\n"
  printf "  -c    enable colors (requires tput)\n"
  printf "  -n    disable colors\n"
  exit "$1"
}

error() {
  printf "${bold}${error_c}Error:${reset} %s.\n" "$2" >&2
  if [ -n "$3" ]; then
    printf "\n%s\n" "$3" >&2
  fi
  exit "$1"
}

download () {
  if available curl; then
    curl "$1" 2>/dev/null; return $?
  elif available wget; then
    wget "$1" -q -O-; return $?
  fi
  error 4 "no downloader installed" \
        "Install curl or wget before proceeding."
}

pager () {
  if [ -n "$DDG_PAGER" ] && available "$DDG_PAGER"; then
    "$DDG_PAGER"; return $?
  elif [ -n "$PAGER" ] && available "$PAGER"; then
    "$PAGER"; return $?
  elif available less; then
    less -r; return $?
  fi
  cat; return $?
}

browser () {
  if [ -n "$DDG_BROWSER" ] && available "$DDG_BROWSER"; then
    "$DDG_BROWSER" "$1"; return $?
  elif [ -n "$BROWSER" ] && available "$BROWSER"; then
    "$BROWSER" "$1"; return $?
  elif available lynx; then
    lynx "$1"; return $?
  elif available w3m; then
    w3m "$1"; return $?
  fi
  error 4 "no browser installed" \
        "Install a browser before proceeding."
}

cache_dir="$HOME/.cache/ddg"

purge_cache () {
  rm -f "${cache_dir}/query"
  rm -f "${cache_dir}/raw"
  rm -f "${cache_dir}/urls"
  rm -f "${cache_dir}/titles"
  rm -f "${cache_dir}/desc"
}

check_age () {
  #TODO
  true
}

check_cache () {
  mkdir -p "${cache_dir}"
  if [ -r "${cache_dir}/query" ]; then
    old_query=$(cat "${cache_dir}/query")
    if [ -z "$query" ] || [ "${query}" = "${old_query}" ]; then
      check_age
      if [ ! -s "${cache_dir}/urls" ]; then
        purge_cache
        error 2 "no cached search results" \
              "Try again with a non-empty QUERY."
      fi
    else
      purge_cache
    fi
  else
    if [ -z "$query" ]; then
      error 2 "no cached search results" \
            "Try again with a non-empty QUERY."
    fi
    purge_cache
  fi
}

clean_up () {
  line_cut=$(grep -n "${cache_dir}/tmp" -e "result-sponsored" | \
    cut -d ':' -f 1 | \
    tail -1)
  [ -z "$line_cut" ] && line_cut=$(grep -n "${cache_dir}/tmp" -e "</style>" | \
    cut -d ':' -f 1)
  tail -n +"${line_cut}" "${cache_dir}/tmp" > "${cache_dir}/raw"
  rm -f "${cache_dir}/tmp"
}

extract_urls () {
  grep "${cache_dir}/raw" -e "result-link" | \
    cut -d '"' -f 4 | \
    cut -d '=' -f 3 | \
    urldecode >> "${cache_dir}/urls"
}

extract_titles () {
  grep "${cache_dir}/raw" -e "result-link" | \
    cut -d ">" -f 2- | \
    split "</a>" 1 | \
    htmlstrip | \
    htmldecode >> "${cache_dir}/titles"
}

extract_descriptions () {
  grep "${cache_dir}/raw" -e "result-snippet" -A 2 | \
    grep -v -e "<td.*>" -e "</td>" -e "--" | \
    sed 's/^ \+//g' | \
    bold | \
    htmldecode >> "${cache_dir}/desc"
}

collate () {
  paste "${cache_dir}/titles" "${cache_dir}/urls" "${cache_dir}/desc"
}

pull_results () {
  results_wanted=$((page * page_size))
  [ "$index" -gt "$results_wanted" ] && results_wanted="$index"
  if [ -r "${cache_dir}/urls" ]; then
    len=$(wc -l "${cache_dir}/urls" | cut -d ' ' -f 1)
  else
    len=0
  fi
  while [ "$len" -lt "$results_wanted" ]; do
    download "https://duckduckgo.com/lite?q=${query}&sc=${len}" > "${cache_dir}/tmp"
    if [ ! $? -eq 0 ]; then
      error 3 "no connection to the internet" \
            "Check your connection and try again."
    fi
    clean_up
    extract_urls
    extract_titles
    extract_descriptions
    len=$(wc -l "${cache_dir}/urls" | cut -d ' ' -f 1)
  done
}

paginate () {
  tail -n +"$(( ( page - 1 ) * page_size + 1 ))" | \
    head -n "$(( page_size ))"
}

do_jump () {
  url=$(tail "${cache_dir}/urls" -n +"$1" | head -1)
  browser "$url"
  exit 0
}

do_urls () {
  paginate < "${cache_dir}/urls"
  exit 0
}

do_dump () {
  collate | paginate
  exit 0
}

do_show () {
  counter="$(( (page - 1) * page_size ))"
  collate | paginate | while read -r line; do
    url="$(echo "$line" | cut -f 2)"
    title="$(echo "$line" | cut -f 1 | fmt -s -w "$width")"
    counter=$((counter + 1))
    echo "$title" | while read -r l; do
      if [ -z "$n" ] ; then
        printf "[%3d] ${bold}${title_c}%s${reset}\n" "$counter" "$l"
      else
        printf "      ${bold}${title_c}%s${reset}\n" "$l"
      fi
      n=$((n + 1))
    done
    printf "      ${url_c}<%s>${reset}\n" "$url"
    if $descriptions; then
      desc="$(echo "$line" | cut -f 3 | fmt -w "$width")"
      echo "$desc" | while read -r l; do
        printf "      %s\n" "$l"
      done
    fi
  done | pager
  exit 0
}

check_query () {
  first_char=$(echo "$query" | cut -c 1)
  if [ "${first_char}" = "!" ]; then
    accel=$(echo "$query" | cut -d ' ' -f 1)
    rest=$(echo "$query" | cut -d ' ' -f 2-)
    case "${accel}" in
      (!g|!google) ./google.sh "${rest}" ;;
      (*) browser "https://duckduckgo.com/lite?q=${query}" ;;
    esac
    exit 0
  fi
}

color () {
  case "$1" in
    (black)   tput setaf 0 ;;
    (red)     tput setaf 1 ;;
    (green)   tput setaf 2 ;;
    (yellow)  tput setaf 3 ;;
    (blue)    tput setaf 4 ;;
    (magenta) tput setaf 5 ;;
    (cyan)    tput setaf 6 ;;
    (white)   tput setaf 7 ;;
  esac
}

init_colors () {
  [ -z "$DDG_TITLE_COLOR" ] && title_c="yellow" || title_c="$DDG_TITLE_COLOR"
  [ -z "$DDG_URL_COLOR" ] && url_c="blue" || url_c="$DDG_URL_COLOR"
  title_c=$(color "$title_c")
  url_c=$(color "$url_c")
  error_c=$(color "red")
  bold=$(tput bold)
  reset=$(tput sgr0)
}

init_vars () {
  cols=80
  available tput && cols=$(tput cols)
  width=$(( cols - 6 - 4 ))
  page_size="$DDG_PAGE_SIZE"
  [ -z "$DDG_PAGE_SIZE" ] && page_size=10
  colors="$DDG_COLORS"
  [ -z "$DDG_COLORS" ] && colors=true
  secret_plan="show"
  descriptions=true
  index=0
  page=1

  error_c=""
  title_c=""
  url_c=""
  bold=""
  reset=""
}

init_vars
while getopts "hlduqg:p:nc" o; do
  case "${o}" in
    (d) secret_plan="dump" ;;
    (l) secret_plan="jump"
        index=1 ;;
    (g) secret_plan="jump"
        index="$OPTARG" ;;
    (p) page="$OPTARG" ;;
    (q) descriptions=false ;;
    (u) secret_plan="urls" ;;
    (n) colors=false ;;
    (c) colors=true ;;
    (h) usage 0 ;;
    (*) usage 1 ;;
  esac
done
shift $((OPTIND - 1))
query=$(printf "$*" | urlencode)
if "$colors" && available tput; then
  init_colors
fi
check_query
check_cache
printf "$query" > "${cache_dir}/query"
pull_results
case "${secret_plan}" in
  (dump) do_dump ;;
  (jump) do_jump "$index" ;;
  (urls) do_urls ;;
  (show) do_show ;;
esac