Replace space, that is not in html tags <> with new line using sed

[matt1311]

Hi, I am working on transforming html code text into the .vert text format. I want to use linux utility sed. I have this regexp which should do the work: s/ \(?![^<>]*>\)/\n/g. I use it like this with sed: echo "you <we try> there" | sed 's/ \(?![^<>]*>\)/\n/g' ... The demanded output should be:
you
<we try>
there
But I get the same string as on input. Is the regexp wrong? Or am I using sed incorrectly? Thanks for your help.

[burschik]

You might wish to refine this a bit, but it basically works:

sed 's/</\n</g; s/>/>\n/g;' input_file

[matt1311]

You might wish to refine this a bit, but it basically works:

sed 's/</\n</g; s/>/>\n/g;' input_file

Damn, this works sometimes, doesn't work for case for example echo "you <we try> there here" |sed 's/</\n</g; s/>/>\n/g;' , the output is:
you
<we try>
there here
Any idea how to deal with this?
Anyway, thanks for answer.

[burschik]

So what exactly do you need? Is every word supposed to be on a line of its own, except for text enclosed in angle brackets, or what?

If you don't mind the empty lines, you could use this:

sed 's/\(<[^>]\+>\| \)/&\n/g;'

And if you do, you can use:

sed 's/\(<[^>]\+>\| \)/&\n/g; s/\n \n/\n/g'

[matt1311]

Yes, every word is supposed to be on a new line, except those in <>. Thanks a lot for your help. This regexp does the work:

sed 's/\(<[^>]\+>\| \)/&\n/g; s/\n \n/\n/g'

Thanks again.