How can this Perl regular expression be expressed in Python?

[John Nagle]

Here's a large Perl regular expression, from a Perl address parser in CPAN:

use re 'eval';
$Addr_Match{street} = qr/
(?:
# special case for addresses like 100 South Street
(?$Addr_Match{direct})\W+ (?{ $_{street} = $^N })
($Addr_Match{type})\b (?{ $_{type} = $^N }))
|
(?$Addr_Match{direct})\W+ (?{ $_{prefix} = $^N }))?
(?:
([^,]+) (?{ $_{street} = $^N })
(?:[^\w,]+($Addr_Match{type})\b (?{ $_{type} = $^N }))
(?:[^\w,]+($Addr_Match{direct})\b (?{ $_{suffix} = $^N }))?
|
([^,]*\d) (?{ $_{street} = $^N })
($Addr_Match{direct})\b (?{ $_{suffix} = $^N })
|
([^,]+?) (?{ $_{street} = $^N })
(?:[^\w,]+($Addr_Match{type})\b (?{ $_{type} = $^N }))?
(?:[^\w,]+($Addr_Match{direct})\b (?{ $_{suffix} = $^N }))?
)
)
/ix;

I'm trying to convert this to Python.

Those entries like "$(Addr_Match{direct}) are other regular expressions,
being used here as subexpressions. Those have already been converted
to forms like "Addr_Match.direct" in Python. But how to call them?
Is that possible in Python, and if so, where is it documented?

John Nagle

 

[Gabriel Genellina]

Here's a large Perl regular expression, from a Perl address parser in
CPAN:

use re 'eval';
$Addr_Match{street} = qr/
(?:
# special case for addresses like 100 South Street
(?$Addr_Match{direct})\W+ (?{ $_{street} = $^N })
($Addr_Match{type})\b (?{ $_{type} = $^N }))
|
(?$Addr_Match{direct})\W+ (?{ $_{prefix} = $^N }))?
(?:
([^,]+) (?{ $_{street} = $^N })
(?:[^\w,]+($Addr_Match{type})\b (?{ $_{type} = $^N }))
(?:[^\w,]+($Addr_Match{direct})\b (?{ $_{suffix} = $^N }))?
|
([^,]*\d) (?{ $_{street} = $^N })
($Addr_Match{direct})\b (?{ $_{suffix} = $^N })
|
([^,]+?) (?{ $_{street} = $^N })
(?:[^\w,]+($Addr_Match{type})\b (?{ $_{type} = $^N }))?
(?:[^\w,]+($Addr_Match{direct})\b (?{ $_{suffix} = $^N }))?
)
)
/ix;

I'm trying to convert this to Python.

Those entries like "$(Addr_Match{direct}) are other regular expressions,
being used here as subexpressions. Those have already been converted
to forms like "Addr_Match.direct" in Python. But how to call them?
Is that possible in Python, and if so, where is it documented?

That would be string interpolation, like this:

Addr_Match = {"direct": "some_re_string",
"type": "other_re"
}

regexp = "%(direct)s %(type)s" % Addr_Match

 

[Gabriel Genellina]

Incidentally, does anybody know what "$^N" means in Perl? That
abbreviation isn't in the list of special variables.

From the context it appears to be the "last matched group", or something
like that... but best look for some authoritative answer.

 

[John Nagle]

Here's a large Perl regular expression, from a Perl address parser in
CPAN:

use re 'eval';
$Addr_Match{street} = qr/
(?:
# special case for addresses like 100 South Street
(?$Addr_Match{direct})\W+ (?{ $_{street} = $^N })
($Addr_Match{type})\b (?{ $_{type} = $^N }))
|
(?$Addr_Match{direct})\W+ (?{ $_{prefix} = $^N }))?
(?:
([^,]+) (?{ $_{street} = $^N })
(?:[^\w,]+($Addr_Match{type})\b (?{ $_{type} = $^N }))
(?:[^\w,]+($Addr_Match{direct})\b (?{ $_{suffix} = $^N
}))?
|
([^,]*\d) (?{ $_{street} = $^N })
($Addr_Match{direct})\b (?{ $_{suffix} = $^N })
|
([^,]+?) (?{ $_{street} = $^N })
(?:[^\w,]+($Addr_Match{type})\b (?{ $_{type} = $^N }))?
(?:[^\w,]+($Addr_Match{direct})\b (?{ $_{suffix} = $^N
}))?
)
)
/ix;

I'm trying to convert this to Python.

Those entries like "$(Addr_Match{direct}) are other regular expressions,
being used here as subexpressions. Those have already been converted
to forms like "Addr_Match.direct" in Python. But how to call them?
Is that possible in Python, and if so, where is it documented?

That would be string interpolation, like this:

Addr_Match = {"direct": "some_re_string",
"type": "other_re"
}

regexp = "%(direct)s %(type)s" % Addr_Match

You're right. I looked at the Perl code, and the strings are just being
inserted, not precompiled as regular expressions and called.

Incidentally, does anybody know what "$^N" means in Perl? That
abbreviation isn't in the list of special variables.

John Nagle