Tuesday 5 August 2008

Importing free-format address lists into vCard (.vcf) and Excel

A colleague had the following problem: he had a Word document containing information about many different companies he dealt with. They were more or less in a standardised form, with the name of the company on the first line of each "record" and the address on the second (or on the same line in some cases), but after that it was pretty haphazard - sometimes telephone numbers were prefixed with TEL:, sometimes not, sometimes the URLs were listed beneath a heading of "Websites:", sometimes not, and so on.

His desire was to export this information in a form that it could be treated much more like a proper database - spreadsheet format would be a start. Exporting the Word document to text was easy, but where to go from there? I started toying around with a Perl script to try to clean up the data, and found that there were some convenient modules around for capturing input to a vCard standard format.

The attached script uses the Text::vCard module (which you'll have to install manually from CPAN as it isn't available using PPM) with its Addressbook and Node packages to build up a whole address book from the input file. Although I have made no attempt at elegance, it has a good stab at parsing UK-style and US-style company names and addresses into their constituent fields, provided they are separated by commas. NB there are no contact names, as the original data file didn't have any - this is left as an exercise for the reader!
#!/usr/bin/perl -w
#
# convert a file to address records in VCF (vCard) format
# reads standard input and writes to standard output
#
if ( $#ARGV + 1 != 0 ) {
print STDERR "usage: parse_addresses <inputfile.txt >outputfile.vcf\n";
exit;
}

use Text::vCard::Addressbook;

my $addressbook = new Text::vCard::Addressbook;
# For testing / debugging: load a pre-existing address book
#my $addressbook =
# Text::vCard::Addressbook->new(
# { 'source_file' => 'C:/temp/Text-vCard-2.03/rfc2426.vcf', } );

while ( !eof STDIN ) {
parseEntry($addressbook);
}

print $addressbook->export();

sub parseEntry {
my ($addressbook) = @_;

my $line = <STDIN>;
chomp($line);

# First line should contain organisation name and optionally its address
my ( $org, $addr );
if ( $line =~ m'^\s*([0-9A-Za-z][^:]*)\:?\s*$' ) {
$org = $1;
chomp( $line = <STDIN> );
$addr = $line;
}
elsif ( $line =~ m'^\s*([0-9A-Za-z][^:]*)\:\s*(\S.*)$' ) {
$org = $1;
$addr = $2;
}
else {
# Not recognised
print STDERR "Not recognised start of entry: $line\n";
while ( ( defined $line ) && ( $line !~ '^\s*$' ) ) {
chomp( $line = <STDIN> );
print STDERR "Discarding: $line\n";
}
return undef;
}

# print STDERR "Organisation: $org -- Address: $addr\n";
my $vCard = $addressbook->add_vcard();
$vCard->version('3.0');
$vCard->add_node({ 'node_type' => 'ORG' })->name($org);
my $adr = $vCard->add_node( { 'node_type' => 'ADR' } );
my @unsorted = ();
$addr =~ s/\W*$//; # Remove trailing spaces and full-stops
foreach my $adrField ( split( ',', $addr ) ) {
if ( $adrField =~ m'^\s*([A-Z][A-Za-z ]+\s)?([-0-9]{4,11})\s*([A-Z][A-Za-z]+)?\s*$' ) {
# US state and zip-code
my $state = $1;
chop( $state ) if defined $state;
my $zip_code = $2;
my $country = $3;
if ( defined $state ) { $adr->region($state); }
$adr->post_code($zip_code);
if ( defined $country ) { $adr->country($country); }
} elsif ( $adrField =~ m'^\s*([A-Z][A-Za-z ]+\s)?([A-Z]{1,2}[0-9O]{1,2}[A-Z]? [0-9O]{1,2}[A-Z]{2})\s*([A-Z][A-Za-z ]+)?\s*$' ) {
# UK county and post-code
my $county = $1;
chop ($county) if defined $county;
my $post_code = $2;
my $country = $3;
if ( defined $county ) { $adr->region($county); }
$post_code =~ s/O([0-9][A-Z]?) ([0-9])/0$1 $2/;
$post_code =~ s/O([A-Z])? ([0-9])/0$1 $2/;
$post_code =~ s/ O([0-9])/ O$1/;
$post_code =~ s/ ([0-9])O([A-Z]{2})/ $1O$2/;
$adr->post_code($post_code);
if ( defined $country ) { $adr->country($country); }
} elsif ( $adrField =~ m'P.*BOX\s*(\d+)'i ) {
# Post Office Box
my $po_box = $1;
$adr->po_box($po_box);
} elsif ( $adrField =~ m'^\s*(\d+\s+[A-Za-z0-9 ]+$)' ) {
# House number and street
my $street = $1;
$adr->street($street);
} elsif ( $adrField =~ m'^\s*([A-Z][A-Za-z0-9 ]+$)' ) {
push @unsorted, $1;
}
}

# Retrieve unsorted items in reverse order
if (!defined $adr->country() && $#unsorted > 2) {
my $country = pop @unsorted;
if (defined $country && $country =~ m'[A-Z][A-Za-z ]+') {
$adr->country( $country );
} else {
push @unsorted, $country;
}
}
if (!defined $adr->region() && $#unsorted > 1) {
my $region = pop @unsorted;
if (defined $region && $region =~ m'[A-Z][A-Za-z ]+') {
$adr->region( $region );
} else {
push @unsorted, $region;
}
}
if (!defined $adr->city()) {
my $city = pop @unsorted;
if (defined $city && $city =~ m'[A-Z][A-Za-z ]+') {
$adr->city( $city );
} else {
push @unsorted, $city;
}
}
if (!defined $adr->street()) {
$adr->street( pop @unsorted );
}
if ($#unsorted >= 0) {
$adr->extended ( join (', ', @unsorted) );
}

chomp( $line = <STDIN> );
while ( ( defined $line ) && ( $line !~ '^\s*$' ) ) {
if ( $line =~ m'^\s*(https?\://\S+)'i ) {
$vCard->url($1);
} elsif ( $line =~ m'(www\.\S+)'i ) {
$vCard->url($1);
} elsif ( $line =~ m'(\S+@\S+)\s*$' ) {
my $email = $1;
my $node = $vCard->add_node( { 'node_type' => 'EMAIL' } );
my @types = qw (work internet);
$node->add_types( \@types );
$node->value($email);
} elsif ( $line =~ m'^\s*(\+?[0-9() ]+)$' ) {
my $tel = $1;
my $node = $vCard->add_node( { 'node_type' => 'TEL' } );
my @types = qw (work voice);
$node->add_types( \@types );
$node->value($tel);
} elsif ( $line =~ m'^\s*tel.*\:\s*(\S.*)$'i ) {
my $tel = $1;
my $node = $vCard->add_node( { 'node_type' => 'TEL' } );
my @types = qw (work voice);
$node->add_types( \@types );
$node->value($tel);
} elsif ( $line =~ m'^\s*fax.*\:\s*(\S.*)$'i ) {
my $tel = $1;
my $node = $vCard->add_node( { 'node_type' => 'TEL' } );
my @types = qw (work fax);
$node->add_types( \@types );
$node->value($tel);
} elsif ( $line =~ m'^[^:]*name[^:]*\:\s*(\S.*)$'i ) {
$vCard->fn($1);
} elsif ( $line =~ m'^\s*([^:]+)\:\s*$' ) {
my $noteHead = $1;
chomp( $line = <STDIN> );
if ( $line =~ m'^\s*(https?\://\S+)'i ) {
$vCard->url($1);
} elsif ( $line =~ m'(www\.\S+)'i ) {
$vCard->url($1);
} elsif ( $line =~ m'(\S+@\S+)\s*$' ) {
my $email = $1;
my $node = $vCard->add_node( { 'node_type' => 'EMAIL' } );
my @types = qw (work internet);
$node->add_types( \@types );
$node->value($email);
} elsif ( $line =~ m'^\s*(\+?[0-9() ]+)$' ) {
my $tel = $1;
my $node = $vCard->add_node( { 'node_type' => 'TEL' } );
my @types = qw (work voice);
$node->add_types( \@types );
$node->value($tel);
} else {
$vCard->note("$noteHead: $line");
}
} elsif ( $line =~ m'^\s*([^:]+)\:\s*(\S.*)$' ) {
my $noteHead = $1;
my $note = $2;
$vCard->note("$noteHead: $note");
} else {
print STDERR "Cannot understand entry detail: $line\n";
}
chomp( $line = <STDIN> );
}
}

1; # End.

(Look out for cut-off long lines in the above - copy/paste works far better in Firefox than in Internet Explorer. If you have problems, ask me to e-mail you a copy of the script). The script spits out the resulting vCard file to the standard output. Anything it can't parse is echoed to standard error.

I found a really nifty converter from VCF to CSV at http://labs.brotherli.ch/vcfconvert/ - once you've got a pile of addresses in spreadsheet format, you can do anything with it.