User:Laurier Rochon/prototyping/pipeline1

1 - Bash file to automate the read using while : sedsongs.sh

while true
do
date | sed -f digits.sed | bash wrap.sh | midge
timidity a.out.mid
sleep 0
done

2 - Bash file rebashing itself : bashbash.sh (thanks Tolga!)

date | sed -f digits.sed | bash wrap.sh | midge
timidity a.out.mid
sh bashbash.sh

3 - Playing a song with ls as an input (grabs all info from ls, puts it in a file and does the same replacement) : ls_songs.sh I guess this could of been even just 1 line... (56 secs playing time)

ls -l > dir.txt

cat dir.txt | sed -f digits.sed | bash wrap.sh | midge
timidity a.out.mid

And then sed file looks a bit different : digits.sed

#keep c-g letters intact
s/[^c-g^0-9]//g

#replace end of lines, tabs, newlines, etc. just in case
s/[ ]*$//g
s/^$//g
s/\n//g
s/\t//g
s/$//g
s/./& /g

#swap numbers...
s/1/c /g
s/2/c+ /g
s/3/d /g
s/4/d+ /g
s/5/e /g
s/6/f+ /g
s/7/g /g
s/8/g+ /g
s/9/a /g
s/0/a+ /g

4. Number 3 makes a really long song with just a few folders since ls seems to spit out a lot of numerical information. To get around it, loop + concatenation. Used previous sed file too. (8 sec playing time)

#file to create
FILE=concat.txt

#check if the file exists already
if [ -f $FILE ];
#if it does, delete and make a new one (the script appends)
then
   echo "File $FILE exists"
   rm $FILE
   echo $FILE "File removed, creating a new one."
else
  echo "File $FILE does not exists, creating it"
fi

for i in *
do
echo $i >> $FILE
done

cat $FILE | sed -f digits.sed | bash wrap.sh | midge
timidity a.out.mid