改了下RavagerAlpha 的从射手获取字幕的shell脚本

[tititake]

地址：

	#!/bin/sh
	# Author: RA <[e-mail]>
	# Author: Chris Yu <[e-mail]>
	USAGE(){
	echo "Usage:$(basename $0) [-eng] [-suf] [-togbk] [-toutf8] files..."
	echo "OPTIONS"
	echo " -eng : download English subtitle, default is Chinese"
	echo " -suf : add language suffix in subtitle file name"
	echo " -togbk : try to convert file content from big5 to gbk with iconv"
	echo " -toutf8 : try to convert file encoding to utf8 with enca"
	}
	[ $# -eq 0 ] && USAGE && exit 0
	ERROR(){
	echo "Error:$@" >&2
	exit 1
	}
	[ -z `which openssl 2>/dev/null` ] || MD5='openssl md5'
	[ -z `which md5 2>/dev/null` ] || MD5='md5'
	[ -z `which md5sum 2>/dev/null` ] || MD5='md5sum'
	[ -z "$MD5" ] && ERROR "No MD5 tools"
	[ -z `which curl` ] && ERROR "No curl"
	[ -z `which dd` ] && ERROR "No dd :)"
	[ -z `which hexdump` ] && ERROR "No hexdump tools"
	[ -z `which gunzip` ] && ERROR "No gunzip tools"
	TMPDIR="/tmp"
	[ -w "$TMPDIR" ] || ERROR "Cannot write to temp dir $TMPDIR"
	USERAGENT="SPlayer Build 1543"
	LANGUAGE='chn'
	TIMEOUT=5
	CRAP="blueray|bluray|blu\-ray|remux|dvdrip|xvid|cd[0-9]|dts|vc1|vc\-1|hdtv|1080p|720p|1080i|x264|limited|ac3|hddvd|repack|@|dts\-hd"
	CRAP_MISC="\[|\]|\.|\-|\#|\_|\=|\+|\<|\>|\,"
	LANG_SUFFIX=""
	TOGBK=0
	TOUTF8=0
	case `uname` in
	'Linux')
	SED_OPTS='-r -e';;
	'FreeBSD')
	SED_OPTS='-E -e'
	TIMEOUT=5000
	;;
	'Darwin')
	# Mac Default , if you using GNU sed,edit it
	SED_OPTS='-E -e';;
	*)
	ERROR "unknown OS";;
	esac
	getFileSize(){
	# Fix Me
	[ -f "$1" ] && ls -nl "$1" | awk '{print $5}' || echo 0
	}
	getFileHash(){
	dd if="$file" bs=1 count=4096 skip=$1 2> /dev/null | $MD5 | head -c32
	}
	stripFileName(){
	echo "$(sed $SED_OPTS "s/"$CRAP"//g" -e "s/"$CRAP_MISC"/ /g")"
	}
	while [ -n "$1" ];do
	[ "$1" = '--help' ] && USAGE && exit 0
	if [ "$1" = '-eng' ]; then
	LANGUAGE='eng'
	shift
	fi
	if [ "$1" = '-suf' ]; then
	LANG_SUFFIX=".${LANGUAGE}"
	shift
	fi
	if [ "$1" = '-togbk' ]; then
	TOGBK=1
	shift
	fi
	if [ "$1" = '-toutf8' ]; then
	TOUTF8=1
	shift
	fi
	oriname="$1"
	file="`realpath \"$1\"`"
	[ ! -f "$file" ] && {
	echo "Cannot locate the target" >&2
	shift
	continue
	}
	filesize=$(getFileSize "$file")
	[ $filesize -le 8192 ] && ERROR "Serious?"
	filepath="E:\\$(dirname "$file" | sed 's/\//\\\\/g')" # it's just work
	filename="$(echo "$(dirname "$oriname")/$(basename "$oriname")" | sed 's/\.[^\.]*$//')"
	moviename="$(basename "$filename" | tr [A-Z] [a-z] | stripFileName | xargs -0)"
	bin="${TMPDIR}/$(basename "$filename").bin"
	first=4096
	second=$(($filesize/3*2))
	third=$(($filesize/3))
	fourth=$(($filesize-8192))
	filehash="$(getFileHash $first);$(getFileHash $second);$(getFileHash $third);$(getFileHash $fourth)"
	serverseq="$(( $RANDOM % 9 + 1 )) $(( $RANDOM % 9 + 1 )) $(( $RANDOM % 9 + 1 ))"
	for i in $serverseq
	do
	SERVER="http://splayer$i.shooter.cn/api/subapi.php"
	echo -n "Sending request to $SERVER..."
	#send request
	curl -s --connect-timeout $TIMEOUT -A "$USERAGENT" \
	-F "pathinfo=$filepath" -F "filehash=$filehash" -F "shortname=$moviename" -F "lang=$LANGUAGE" \
	-o "$bin" "$SERVER"
	# sometimes shooter.cn give zero fuck about your request , so try other
	# suck code
	if [ $? -eq 0 -a $(getFileSize "$bin") -gt 1024 ]; then
	echo "OK,Extracting"
	FLAG="DONE"
	else
	[ -f "$bin" ] && rm "$bin"
	echo "Fail" >&2
	FLAG="FAIL"
	sleep 2
	continue
	fi
	# get the subtitle filetype, srt ass ssa or smi
	matches=`grep -aboU -E "srt|ass|ssa|smi|sub" "$bin"`
	validmatches=""
	validcount=0
	for extgrep in $matches; do
	extpos=${extgrep%:*}
	extname="${extgrep##*:}"
	matched=1
	if [ $validcount -gt 0 ] ; then
	extflagpos=$(( $extpos - 4 ))
	extflag=`dd if="$bin" bs=1 count=8 skip=$extflagpos 2> /dev/null | hexdump -x |awk '{print $2$3}' 2> /dev/null`
	if [ "$extflag" = '00000300' ] || [ "$extflag" = '00000003' ] ;then
	subendpos=$(( $extpos - 21 ))
	validmatches="$validmatches$subendpos "
	else
	matched=0
	fi
	fi
	if [ $matched -eq 1 ] ; then
	subbeginpos=$(( $extpos + 7 ))
	validmatches="$validmatches$extname-$subbeginpos-"
	validcount=$(( $validcount + 1 ))
	fi
	done
	if [ $validcount == 0 ];then
	FLAG="FAIL"
	echo "No subtitle found, shouldn't happend!"
	break;
	else
	echo "Got $validcount subtitle(s)"
	fi
	subcount=0;
	for ii in $validmatches; do
	extname=$(echo $ii | cut -f1 -d-)
	subbegin=$(echo $ii | cut -f2 -d-)
	subend=$(echo $ii | cut -f3 -d-)
	#echo "$extname , $subbegin, $subend . "
	sub="${oriname%.*}${LANG_SUFFIX}.$extname"
	if [ $subcount -gt 0 ]; then
	sub="${oriname%.*}.${LANGUAGE}${subcount}.$extname"
	fi
	if [ "$subend" == "" ];then
	dd if="$bin" of="$sub" bs=1 skip=$subbegin 2> /dev/null
	else
	count=$(( $subend - $subbegin ))
	dd if="$bin" of="$sub" bs=1 skip=$subbegin count=$count 2> /dev/null
	fi
	# maybe not handle well , fix me
	gzip=`dd if="$sub" bs=1 count=2 2> /dev/null | hexdump -x | awk '{print $2}'`
	([ "$gzip" = '8b1f' ] || [ "$gzip" = '1f8b' ]) && {
	#echo "Got gzip here,unzipping..."
	mv "$sub" "${sub}.tgz"
	gunzip -c "${sub}.tgz" > "$sub" 2>/dev/null
	rm "${sub}.tgz"
	}
	subcount=$(( $subcount + 1 ))
	echo "#$subcount : [$(basename "$sub")]"
	# try to convert to gbk and utf-8
	if [ $TOGBK == 1 ] && [ ! -z `which iconv 2>/dev/null` ]; then
	iconv -f big5 -t gbk "$sub" -o "$sub".tmp_iconv 2>/dev/null \
	&& mv "$sub".tmp_iconv "$sub"
	rm "$sub".tmp_iconv 2>/dev/null
	fi
	if [ $TOUTF8 == 1 ] && [ ! -z `which enca 2>/dev/null` ]; then
	enca -L zh_CN -x utf-8 "$sub" 2>/dev/null
	fi
	done
	rm "$bin"
	break
	done
	[ "$FLAG" = "FAIL" ] && echo "Cannot find the subtitle:[$(basename "$file")]" >&2
	shift
	done

view raw fetchsub.sh hosted with ❤ by GitHub

原作者页面在这里 http://www.v2ex.com/t/51077 。
先感谢下原作者 @RavagerAlpha

主要运行在我Linux的NAS Server上。

改动的内容
1. 判断字幕扩展名和解出字幕文件部分，增强了兼容性，只支持 srt/ass/ssa，依赖grep。
2. gzip判断时考虑高低位位置颠倒的问题。
3. 存在enca时自动转字幕文件为utf-8格式。
4. 如果参数是视频文件的软链接，字幕保存在链接文件所在目录，命名以链接文件为准。

水平有限，欢迎改进。

[jjgod]

为什么 v2ex 贴 gist 格式全是错的？

[11138]

@Livid

[duhastmich]

以前用过一个submarine, 为什么会有 "E:"

[tititake]

@duhastmich 那个是传到服务器的一个参数，就是为了拼成一个Win形式的绝对路径，所以C: D: E:都无所谓。