Replaces unicode characters with ASCII-safe alternatives

findreplace.sh

#!/bin/bash

# findreplace
# ===========
# Replaces unicode characters with ASCII-safe alternatives
# Single-letter words á, à and é are replaced with a', 'a and e'

## check commands
SED=sed
if gsed --version >/dev/null 2>&1 ; then
    SED=gsed
elif ! sed --version >/dev/null 2>&1 ; then
    echo "GNU sed is required."
    exit 0
fi

## usage info
if [[ $# -eq 0 ]] ; then
cat << EOF
Usage:
    $0 <find arguments>
Example:
    $0 . -name '*.java'
EOF
    exit 0
fi

## regex
read -r -d '' REGEX << EOF
    s/\bà\b/\'a/g;
    s/\bá\b/a\'/g;
    s/\bé\b/e\'/g;

    s/[áàâã]/a/g;
    s/[éê]/e/g;
    s/í/i/g;
    s/[óôõ]/o/g;
    s/ú/u/g;
    s/ç/c/g;
EOF
# upper case
REGEX=$REGEX$(echo $REGEX | $SED "s/[^sgb]*/\U&/g")

## parallel commands
read -r -d '' COMMANDS << EOF
    echo "{}"
    $SED -i "$REGEX" "{}"
    $SED '/\([\d000-\d127]\)/d' "{}"
EOF

## run
find "$@" -print0 | parallel --will-cite -0 "$COMMANDS"