sanchar

#!/usr/bin/env bash
# Get text from clipboard
# and transform certain characters to simpler ones
# and write the result back to the clipboard.
#
# Alternatively, sanitize given text files directly.
#
# Usage:
#
#   sanchar                  # copy sanitized clipboard text back to clipboard
#   sanchar FILE [FILE...]   # sanitize file(s) in-place
#
# Beware: editing some lines in this script in some editors may lead to quirky results, e.g. in vim,
# which apparently has issues coping with some of the invisible characters.
#
# author: andreasl

replace_chars() {
    sed \
        -e "s/[’‘‚‛]/'/g" \
        -e 's/[“”„＂]/"/g' \
        -e 's/[–—−‑‒―]/-/g' \
        -e 's/[…]/.../g' \
        -e 's/[•‣◦▪●]/-/g' \
        -e 's/[×]/x/g' \
        -e 's|[÷]|/|g' \
        -e 's/[，]/,/g' \
        -e 's/[．]/./g' \
        -e 's/[！]/!/g' \
        -e 's/[？]/?/g' \
        -e 's/[；]/;/g' \
        -e 's/[～]/~/g' \
        -e 's/[（]/(/g' \
        -e 's/[）]/)/g' \
        -e 's/[［]/[/g' \
        -e 's/[］]/]/g' \
        -e 's/[：]/:/g' \
        -e 's/[｜]/|/g' \
        -e 's|[⧸⁄∕]|/|g' \
        -e 's|[＜]|<|g' \
        -e 's|[＞]|>|g' \
        -e 's/[             　]/ /g' \
        -e 's/[­​‌‍‎‏‪‫‬‭‮⁠]//g' # some invisible characters
}

remove_leading_and_trailing_blank_lines() {
    # Remove leading and trailing blank lines, but keep the ones in the middle.
    # Blank lines are not printed right away, they are held back until a line with content
    # shows up and releases them. Leading blanks are dropped because no content has been seen
    # yet, trailing ones because the input ends before anything releases them. The blanks in
    # between survive, because the next content line flushes them.
    awk '
        /[^[:space:]]/ {                  # line has at least one non-whitespace character
            printf "%s", blanks           # print the blank lines collected in `blanks`
            blanks = ""                   # reset `blanks`
            seen = 1                      # remember that content has appeared
            print                         # print this line
            next                          # done with this line, skip the rule below
        }
        seen { blanks = blanks $0 "\n" }  # blank line after content: hold it, do not print
    '
}

show_help() {
    cat <<'EOF'
Usage:
  sanchar                  sanitize clipboard text and write back to clipboard
  sanchar FILE [FILE...]   sanitize file(s) in-place

Options:
  -h, --help  show this help message and exit
EOF
    exit 0
}

case "${1:-}" in
-h | --help) show_help ;;
esac

if [ $# -eq 0 ]; then
    command -v pbpaste >/dev/null && clip_out=(pbpaste) || clip_out=(xclip -selection clipboard -out)
    command -v pbcopy >/dev/null && clip_in=(pbcopy) || clip_in=(xclip -selection clipboard)

    # sanitize clipboard
    input_text="$("${clip_out[@]}")"
    if [ -z "$input_text" ]; then
        echo "Clipboard is empty or does not contain text."
        exit 1
    fi

    output_text="$(echo "$input_text" | replace_chars | remove_leading_and_trailing_blank_lines)"

    printf '%s' "$output_text" | "${clip_in[@]}"

    echo "Special characters have been replaced and the clipboard has been updated."
else
    # sanitize given files
    for file in "$@"; do
        # add a sentinel `x` character to presever trailing newlines
        input_text="$(
            cat "$file"
            printf x
        )"
        input_text="${input_text%x}"
        tmp="$(mktemp)"
        printf '%s' "$input_text" | replace_chars >"$tmp"
        mv "$tmp" "$file"
        echo "Sanitized: $file"
    done
fi