ATENÇÃO: Este conteúdo foi publicado há 5 anos. Eu talvez nem concorde mais com ele. Se é um post sobre tecnologia, talvez não faça mais sentido. Mantenho neste blog o que escrevo desde os 14 anos por motivos históricos. Leia levando isso em conta.
Suppose you have a lot of .doc, .docx, .xls, .xlsx, .gz, .bz2, .pdf and text in general (.csv, .txt etc.) files and want to dump all the (unique) email addresses from them. How would you do it? Here is a simple solution I’ve just implemented (and probably didn’t test enough, so tell me if you find any bug):
#!/bin/sh tmp=$(tempfile) while [ $# -gt 0 ]; do if [ -r "$1" ]; then ext=$(echo ${1#*.} | tr [A-Z] [a-z]) case $ext in docx | xlsx) # requires: http://blog.kiddaland.net/2009/07/antiword-for-office-2007/ cat_open_xml "$1" >> $tmp ;; doc) # requires: antiword antiword "$1" >> $tmp ;; xls) # requires: catdoc xls2csv "$1" >> $tmp ;; gz) cat "$1" | gunzip >> $tmp ;; bz2) cat "$1" | bunzip2 >> $tmp ;; zip) unzip -p "$1" >> $tmp ;; pdf) # requires: xpdf-utils t=$(tempfile) pdftotext "$1" $t cat $t >> $tmp rm $t ;; *) text=$(file -b --mime-type "$1" | sed -e 's//.*//') if [ "z$text" = "ztext" ]; then cat "$1" >> $tmp fi ;; esac fi shift done cat $tmp | grep -o -E '\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+.[A-Za-z]{2,4}\b' | tr [A-Z] [a-z] | sort -u rm $tmp |
(the email regexp is explained here: regular-expressions.info/email.html)