#!/usr/bin/env bash
echo "[*] Script started"
input="pdf_list.txt"
keyword=$1
output=$2

echo "[*] Finding all possible pdf files"
find . -type f -name "*.pdf" > "$input"

echo "[*] PDF files found and saved in '"$input"'"
cat "$input"

echo "[*] Finding possible emails"
while IFS= read -r line
do

    pdftotext "$line" - | grep "$keyword" >> tmp_list.txt
    grep -i -o '[A-Z0-9._%+-]\+@[A-Z0-9.-]\+\.[A-Z]\{2,4\}' tmp_list.txt >> tmp_final.txt
    #sort -u tmp_final.txt > final_list.txt
    sort -u tmp_final.txt > "$output"

done < "$input"

echo "[*] Emails found and saved in $output"
echo "[*] Cleaning up"
rm -f tmp_list.txt && rm -f tmp_final.txt && rm -f "$input"
echo "[*] Done"