# (may be /bin/bash or perhaps any default shell)

# This chain of sed scripts creates a *.srt file, containing subtitles 
# in the second language only, 
# using a source file with two languages, formatted (with no blank lines):

# mm:ss
# original language
# second language (translation of the original one)

# A source text formatted 

# mm:ss 
# original language subtitle

# is easy to copy from youtube, 
# provided the manually-written subtitles are there 
# - use the transcript icon, the 4th option below the video.

# (by Jenik Hollan, CzechGlobe, May 2014)

sed -r "{N; N; s/\n[^\n]+\n/,000 /; s/^/00:/} " en_cz.txt | \
sed -r "=" | \
sed -r "{N; s/^([0-9]+)\n(.+)/\1 \2/}" | \
sed -r "{s/^([0-9]+) ([0-9:,]+) (.+)/\2\n\1 \2 \3/}" | \
sed -r '1d; ${s/^([0-9]+) ([0-9:,]+) (.+)/\1 \2 \3\n\2/}' | \
sed -r '${s/,000/,900/}' | \
sed -r "
{s/,000/,100/; N; s/([0-9]+) ([0-9:]+,100) ([^\n]*)\n(.+)$/\n\1\n\2 --> \4\n\3/}
" > cz.srt