Question

Using Ruby, I’m trying to parse some documentation in which I need to split blocks of text, each with a heading and followed by an unknown length of text, and push them to an array;

SECTION 1. A HEADING

Some undetermined length of text,
which can be multiple lines and paragraphs.


SECTION 2. ANOTHER HEADING

Another big block of text.

should become

["SECTION 1. A HEADING

Some undetermined length of text,
which can be multiple lines and paragraphs.",
"SECTION 2. ANOTHER HEADING

Another big block of text."]

I could just use string.split(/\n\n\n/), but I want something more specific as I can’t guarantee that each section will have two blank lines after it. A little more experimenting led me to this;

string.split(/(?:^|\n)(SECTION.+\n)/).each do |s|
  sections << s
end

but I’d have to process the output again to get what I need.

Is there some way to get this done without having to do multiple passes?

Thanks for your help.

Answer 1

You can use String#scan with multiline-mode regexp and positive look-ahead:

text = <<ENDTEXT
SECTION 1. A HEADING

Some undetermined length of text,
which can be multiple lines and paragraphs.


SECTION 2. ANOTHER HEADING

Another big block of text.
ENDTEXT

header = /^SECTION\s+\d+\./
sections = text.scan(/(?m)#{header}.*?(?=#{header}|\Z)/)

puts sections.join("\n---\n")

# =>
SECTION 1. A HEADING

Some undetermined length of text,
which can be multiple lines and paragraphs.



---
SECTION 2. ANOTHER HEADING

Another big block of text.