Improve get images from internet script (via DuckDuckGo)

scripts/01_download_images.sh

 #!/usr/bin/env bash
 
+set -o errexit
+set -o nounset
+set -o pipefail
+
 command -v convert >/dev/null 2>&1 || { echo >&2 "I require convert (imagemagick) but it's not installed. Aborting."; exit 1; }
 command -v jq >/dev/null 2>&1 || { echo >&2 "I require jq (json parser) but it's not installed. Aborting."; exit 1; }
 
 CURRENT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" >/dev/null 2>&1 && pwd)"
 BASE_DIR="$(dirname "${CURRENT_DIR}")"
 
-IMAGES_CACHE_FOLDER="${CURRENT_DIR}/cache/download"
-INPUT_WORDS_LIST="${BASE_DIR}/assets/files/words.json"
+# CSV source file for words
+SOURCE_CSV_FILE="${CURRENT_DIR}/words.csv"
+MAX_RANDOM_WORDS_TO_DOWNLOAD=100
+MAX_IMAGES_PER_WORD=5
 
 # Images variants
 KEYWORD_VARIANTS=",image,picture,drawing,black and white,painting,icon"
-COUNT_BY_WEYWORD_VARIANT=3
+
+IMAGES_CACHE_FOLDER="${CURRENT_DIR}/cache/download"
+mkdir -p "${IMAGES_CACHE_FOLDER}"
+
 KEYWORD_VARIANTS_LIST="$(echo "${KEYWORD_VARIANTS}" | tr "," "\n")"
 
-# QWANT parameters
-LANG="fr_FR"
-TYPE="images"
-OFFSET=0
-BASE_URL="https://api.qwant.com/api/search/${TYPE}"
+get_images_from_ddg() {
+  local -r QUERY_STRING="$1"
+  local -r BASE_URL="https://duckduckgo.com"
+
+  local -r TOKEN="$(curl --silent "${BASE_URL}/q=${QUERY_STRING}" | sed 's/[&,]/\n/g' | sed 's/"//g' | grep "vqd" | tail -n 1 | cut -d'=' -f2)"
+
+  local -r SEARCH_URL="https://duckduckgo.com/?t=ffab&q=${QUERY_STRING}&atb=v301-1&iax=images&ia=images"
+  local -r RESULT="$(curl --silent "${SEARCH_URL}")"
+  local -r JSON_URL="${BASE_URL}$(echo "${RESULT}" | sed 's/ /\n/g' | grep "initialize" | cut -d';' -f2 | cut -d "'" -f2 | sed 's/\/d\.js/\/i.js/g')"
+
+  local -r RESULTS="$(curl --silent "${JSON_URL}")"
+  local -r THUMBNAILS_URLS="$(echo "${RESULTS}" | sed 's/ /\n/g' | sed 's/,/\n/g' | grep '"thumbnail":' | cut -d '"' -f4 | head -n ${MAX_IMAGES_PER_WORD})"
+  echo "${THUMBNAILS_URLS}"
+}
 
-WORDS_LIST="$(cat "${INPUT_WORDS_LIST}" | grep -E '\"fr\": \"[A-Z\-]+\",$' | cut -d'"' -f4 | sort | uniq | sort -R)"
+WORDS_LIST="$(cat "${SOURCE_CSV_FILE}" | cut -d';' -f1 | sort | uniq | sort -R | head -n ${MAX_RANDOM_WORDS_TO_DOWNLOAD})"
 while read -r KEYWORD; do
   if [[ -n "${KEYWORD}" ]]; then
     echo "KEYWORD: ${KEYWORD}"
@@ -33,17 +51,7 @@ while read -r KEYWORD; do
         QUERY_STRING="${QUERY_STRING} ${VARIANT}"
       fi
 
-      # Get QWANT API query from keyword
-      QUERY="$(echo "${QUERY_STRING}" | tr "A-Z" "a-z" | sed 's| |%20|g')"
-      QUERY_URL="${BASE_URL}?count=${COUNT_BY_WEYWORD_VARIANT}&q=${QUERY}&t=${TYPE}&safesearch=1&offset=${OFFSET}&locale=${LANG}&uiv=4"
-      # echo "QUERY_URL: ${QUERY_URL}"
-
-      # Get QWANT thumbnails urls from keyword
-      URL_LIST="$(curl "${QUERY_URL}" \
-          --silent \
-          -H 'User-Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:87.0) Gecko/20100101 Firefox/87.0' \
-          --compressed \
-          | jq | grep "media_preview" | cut -d'"' -f4 | sed 's|^//||g')"
+      URL_LIST="$(get_images_from_ddg "${QUERY_STRING}")"
 
       if [[ -z "${URL_LIST}" ]]; then
         echo "    No image found..."