ambevar-dotfiles/.scripts/translate

#!/bin/sh

## TODO: support for long texts.

_usage () {
	cat<<EOF
Usage: ${1##*/} [-i source-language] target-language [text]

Translate text from one language to another.  The languages are specified using
the international prefix (e.g. 'en' for English, 'de' for German, etc.).

When omitting the source language, the service tries to guess the language from
the input content.

When omitting the text, standard input is read.

Options:

  -h: Show this help.
  -i: Input language (optional).
  -l: Show alternate translations (single words only). If jshon is found on the
      system, print all translations for all gramatical types (noun, verb,
      etc.).

Language codes:

af: Afrikaans, sq: Albanian, ar: Arabic, hy: Armenian, az: Azerbaijani, eu:
Basque, be: Belarusian, bn: Bengali, bs: Bosnian, bg: Bulgarian, ca: Catalan,
ceb: Cebuano, zh-CN, Chinese (Simplified) zh-TW, Chinese (Traditional) hr:
Croatian, cs: Czech, da: Danish, nl: Dutch, en: English, eo: Esperanto, et:
Estonian, tl: Filipino, fi: Finnish, fr: French, gl: Galician, ka: Georgian, de:
German, el: Greek, gu: Gujarati, ht: Haitian, Creole iw: Hebrew, hi: Hindi, Hmn:
Hmong, hu: Hungarian, is: Icelandic, id: Indonesian, ga: Irish, it: Italian, ja:
Japanese, jw: Javanese, kn: Kannada, km: Khmer, ko: Korean, lo: Lao, la: Latin,
lv: Latvian, lt: Lithuanian, mk: Macedonian, ms: Malay, mt: Maltese, mr:
Marathi, no: Norwegian, fa: Persian, pl: Ryba, pt: Portuguese, ro: Romanian, ru:
Russian, sr: Serbian, sk: Slovak, sl: Slovenian, es: Spanish, sw: Swahili, sv:
Swedish, ta: Tamil, te: Telugu, th: Thai, tr: Turkish, uk: Ukrainian, ur: Urdu,
vi: Vietnamese, cy: Welsh, yi: Yiddish

EOF
}

unset SL
OPT_LONG=false
while getopts ":hi:l" opt; do
	case $opt in
	h)
		_usage "$0"
		exit 1
		;;
	i)
		SL="$OPTARG" ;;
	l)
		OPT_LONG=true ;;
	\?)
		_usage "$0"
		exit 1
		;;
	esac
done

shift $(($OPTIND - 1))

if [ $# -eq 0 ]; then
	_usage "$0"
	exit 1
fi

TL=$1

shift

_filter () {
	if ! $OPT_LONG; then
		sed -e 's/\[\[\["\([^"]*\).*/\1/' -e 's/"//g'
		echo
		return
	fi

	if command -v jshon >/dev/null 2>&1; then
		## Input may have empty fields, which is not normally allowed in
		## json. Let's fix this.
		input="$(sed -e 's/,\+/,/g' -e 's/\[,\+/\[/g' -e 's/,\+\]/\]/g')"

		## The use of jshon without checking the array size is unsafe. If the
		## size is not as expected, it probably means that source and target
		## languages are the same. Anyhow, we drop the error output.

		## Google Translate does not have the same structure when English is
		## involved. With English, the list of translations is in cell 1,
		## whitout English it is in cell 3. Without English, there is no
		## distinction on grammatical type.
		if [ "$(echo "$input" | jshon -e 1 -t)" = "string" ]; then
			## No English
			echo "$input" | jshon -e 3 -e 0 -e 2 -a -e 0 -u | tr '\n' ',' | sed 's/,$/\n/'
		else
			count=0
			## Fetch grammatical type in 'type'.
			while read -r type; do
				echo -n $type:

				## Fetch translation alternative from count-th entry
				echo "$input" | jshon -e 1 -e $count -e 1 -a -u | tr '\n' ',' | sed 's/,$/\n/'
				count=$(($count+1))

			done<<EOF
$(echo "$input" | jshon -e 1 -a -e 0 -u)
EOF
		fi 2>/dev/null

	else
		## Use sed to fetch first entry only.
		sed -e 's/\[\[\[[^[]*\[\[[^[]*\[\([^[]*\)\].*/\1/' -e 's/"//g'
		echo
	fi
}

_translate () {
	TEXT="$(tr '\n' ' ' | sed 's/[\t ]\+/%20/g')"
	## If input is not a single word, disable alternative translations. We use a
	## point in the grep string to ignore any space at the end.
	[ -n "$(echo $TEXT | grep '%20.')" ] && OPT_LONG=false

	if [ "$SL" = "" ]; then
		curl -A "Mozilla/5.0" 'http://translate.google.com/translate_a/t?client=t&text='$TEXT'&tl='$TL'&ie=UTF-8&oe=UTF-8'
	else
		curl -A "Mozilla/5.0" 'http://translate.google.com/translate_a/t?client=t&text='$TEXT'&sl='$SL'&tl='$TL'&ie=UTF-8&oe=UTF-8'
	fi 2>/dev/null | _filter
}

if [ $# -ne 0 ]; then
	echo "$@" | _translate
else
	_translate
fi

## Original command:
# curl -A "Mozilla/5.0" 'http://translate.google.com/translate_a/t?client=t&text=hello&hl=en&sl=en&tl=zh-CN&ie=UTF-8&oe=UTF-8&multires=1&prev=btn&ssel=0&tsel=0&sc=1' | sed 's/\[\[\["\([^"]*\).*/\1/'