Regex for parsing figure references - php

Question

I'm trying to create a regex to parse figures references inside a text. I must match at least these cases:
Fig* 1, 2 and 3 (not only 3, any number)
Fig* 1-3
Fig* 1 and 2
Fig* 1
Fig* 1 to 4
So I tried the following regex:
(Fig[a-zA-Z.]*)(\s(\d(,|\s)* )+|\d\s|and\s\d|\s\d-\d|\s\d)*
The best result would be having the numbers separated, but having the match I can just clean up the result and parse the numbers.
But I just can't seem to be able to parse that "1 to 4". Also, this regex seems not optmized at all. Any ideas?
Here is a sample: http://www.phpliveregex.com/p/3Zj

Answer 1

try this:
(Fig.*) ((\d( to | and |-)\d)|\d)|(\d,\d and \d)

Answer 2

You can use this pattern:
(Fig(?:ures?|s\.)) (\d+(?:(?:-|, | (?:and|to) )\d+)*)
If you need more flexibility, you can replace spaces with \h+ or \h*

Answer 3

edit:
I see my previous regex didn't work.
Atempting redemption, I offer two alternatives that do work -
1.
Using Multi-Line mode - This uses the \G anchor which provides a means
to get an aligned and trimmed output suitable for array
# '/(^Fig[a-zA-Z.]*\h+|(?!^)\G)(?(?<=\d)\h*,\h*)(\d+)(?|\h*(-)\h*(\d+)|\h+(and)\h+(\d+)|\h+(to)\h+(\d+))?/'
( # (1 start)
^ Fig [a-zA-Z.]* \h+ # Fig's
| # or,
(?! ^ ) # Start at the end of last match
\G
) # (1 end)
(?(?<= \d ) # Conditional, if previous digit
\h* , \h* # Require a comma
) # End conditional
( \d+ ) # (2), Digit
(?| # Branch reset (optionally, one of the (-|and|to) \d forms)
\h*
( - ) # (3), '-'
\h*
( \d+ ) # (4), Digit
| \h+
( and ) # (3), 'and'
\h+
( \d+ ) # (4), Digit
| \h+
( to ) # (3), 'to'
\h+
( \d+ ) # (4), Digit
)?
Perl test case
$/ = undef;
$str = <DATA>;
while ($str =~ /(^Fig[a-zA-Z.]*\h+|(?!^)\G)(?(?<=\d)\h*,\h*)(\d+)(?|\h*(-)\h*(\d+)|\h+(and)\h+(\d+)|\h+(to)\h+(\d+))?/mg)
{
length($1) ?
print "'$1'\t'$2'\t'$3'\t'$4'\n" :
print "'$1'\t\t'$2'\t'$3'\t'$4'\n" ;
}
__DATA__
Figs. 1, 2, 3 and 4
Figures 1, 2
Figs. 1 and 2
Figure 1-3
Figure 1 to 3
Figure 1
Output >>
'Figs. ' '1' '' ''
'' '2' '' ''
'' '3' 'and' '4'
'Figures ' '1' '' ''
'' '2' '' ''
'Figs. ' '1' 'and' '2'
'Figure ' '1' '-' '3'
'Figure ' '1' 'to' '3'
'Figure ' '1' '' ''
2. Using Multi-Line mode - This matches entire line, where capture group 1 contains 'Figs',
group 2 contains all the number forms
# '/^(Fig[a-zA-Z.]*\h+)((?(?<=\d)\h*,\h*|\d+(?:\h*-\h*\d+|\h+and\h+\d+|\h+to\h+\d+)?)+)\h*$/'
^
( Fig [a-zA-Z.]* \h+ ) # (1), Fig's
( # (2 start), All the num's
(?(?<= \d ) # Conditional, if previous digit
\h* , \h* # Require a comma
| # or
\d+ # Require a digit
(?: # (and optionally, one of the \d (-|and|to) \d forms)
\h* - \h* \d+
| \h+ and \h+ \d+
| \h+ to \h+ \d+
)?
)+ # End conditional, do many times
) # (2 end)
\h*
$