another regex puzzle - detect capitalised phrases

[cybmole]

I suspect there is no easy answer to this but I will ask anyway.

given a book which uses capitalisation in lieu of scene breaks, with all paragraphs sharing the same CSS i.e.

THIS IS HOW THE 1st paragraphs starts.......blah blah
but not the next paragraph...
Or the one after that......
....
YET SOME TIME LATER THERE is another instance
...

I want to pick out those capitalised starts in order to assign a unique CSS class.

but devising a rule is very hard.

testing that 2nd letter of a paragraph is capitalised works most times but will miss
I CANNOT GET THIS one... and will miss A TOUGH ACT TO follow
and will mis-classify
"I don't want this one"

any better methods, anyone ?

[WS64]

<p>[^a-z]{4,}

Will find any paragraph with the first 4 being no lower ones.
Of course this still will find <p>U.S.A. is the country...
I guess you need to check if the 4 is enough.

[mmat1]

Originally Posted by cybmole

Phrase 1: THIS IS HOW THE
Phrase 2: A TOUGH ACT TO follow

<p>([A-Z])([A-Z|\s[A-Z])

match both.

Actually

<p>([A-Z])([\sA-Z])

will work as well

[theducks]

Originally Posted by cybmole

I suspect there is no easy answer to this but I will ask anyway.

given a book which uses capitalisation in lieu of scene breaks, with all paragraphs sharing the same CSS i.e.

THIS IS HOW THE 1st paragraphs starts.......blah blah
but not the next paragraph...
Or the one after that......
....
YET SOME TIME LATER THERE is another instance
...

I want to pick out those capitalised starts in order to assign a unique CSS class.

but devising a rule is very hard.

testing that 2nd letter of a paragraph is capitalised works most times but will miss
I CANNOT GET THIS one... and will miss A TOUGH ACT TO follow
and will mis-classify
"I don't want this one"

any better methods, anyone ?

([A-Z]* ){2,}

Case sensitive
find 1 or more Upper followed by a space, 2 or more times

[cybmole]

thanks for all the different solutions - you guys make it look so easy !

[Perkin]

Originally Posted by theducks

([A-Z]* ){2,}

Case sensitive
find 1 or more Upper followed by a space, 2 or more times

Shouldn't that be a + instead of the *, as * is 0 or more times, which would match paragraphs that have several spaces at beginning, which are quite often (badly) used for indents.

[theducks]

Originally Posted by Perkin

Shouldn't that be a + instead of the *, as * is 0 or more times, which would match paragraphs that have several spaces at beginning, which are quite often (badly) used for indents.

Yes
(I so rarely use *, so why would I type it here )