--- trunk/all2xml.pl 2003/02/16 22:41:37 13
+++ trunk/all2xml.pl 2003/03/15 21:48:48 42
@@ -8,6 +8,7 @@
use Text::Unaccent 1.02; # 1.01 won't compile on my platform,
use Text::Iconv;
use Config::IniFiles;
+use Encode;
$|=1;
@@ -33,19 +34,23 @@
my $db_dir;
-#die "usage: $0 -d [database_dir] -m [database1,database2] " if (! %opts);
+Text::Iconv->raise_error(1); # Conversion errors raise exceptions
-#print Dumper($config->{indexer});
-#print "-" x 70,"\n";
+# this is encoding of all files on disk, including import_xml/*.xml file and
+# filter/*.pm files! It will be used to store strings in perl internally!
+my $codepage = 'ISO-8859-2';
+
+my $utf2cp = Text::Iconv->new('UTF-8',$codepage);
+# this function will convert data from XML files to local encoding
+sub x {
+ return $utf2cp->convert($_[0]);
+}
-Text::Iconv->raise_error(1); # Conversion errors raise exceptions
+# decode isis import codepage
+my $isis2cp;
-#my $isis_codepage = Text::Iconv->new($config->{isis_codepage},'UTF8');
-#my $index_codepage = Text::Iconv->new($config->{isis_codepage},$config->{index_codepage});
-my $isis_codepage;
-my $index_codepage;
-my $cludge_codepage = Text::Iconv->new('UTF8','ISO8859-1');
-my $xml_codepage;
+# outgoing xml must be in UTF-8
+my $cp2utf = Text::Iconv->new($codepage,'UTF-8');
sub isis2xml {
@@ -62,46 +67,104 @@
my %field_usage; # counter for usage of each field
- foreach my $field (keys %{$config->{indexer}}) {
+ # sort subrouting using order="" attribute
+ sub by_order {
+ return 0 if (! $config->{indexer}->{$a}->{order});
+ return 0 if (! $config->{indexer}->{$b}->{order});
+
+ return $config->{indexer}->{$a}->{order} <=>
+ $config->{indexer}->{$b}->{order} ;
+ }
+
+ foreach my $field (sort by_order keys %{$config->{indexer}}) {
+
+ $field=x($field);
$field_usage{$field}++;
my $swish_data = "";
my $display_data = "";
- my $index_data = "";
+ my $line_delimiter;
+
+ my ($swish,$display);
foreach my $x (@{$config->{indexer}->{$field}->{isis}}) {
- my $format = $x->{content};
+ my $format = x($x->{content});
+ my $delimiter = x($x->{delimiter}) || ' ';
+
+ my $isis_i = 0; # isis repeatable offset
+
my ($s,$d,$i) = (1,1,0); # swish, display default
$s = 0 if (lc($x->{type}) eq "display");
$d = 0 if (lc($x->{type}) eq "swish");
($s,$d,$i) = (0,0,1) if (lc($x->{type}) eq "index");
-#print STDERR "## s: $s d: $d i: $i ## $format ##\n";
- $format = $cludge_codepage->convert($format);
- my ($swish,$display) = parse_format($format,$row);
-#print STDERR "s: $swish\nd: $display\n" if ($swish);
-
-#print STDERR "swish: $swish<-- display: $display<--\n";
- # FIX: this is ugly, UGLY, cludge: OpenIsis return
- # UTF8 encoding of strings, but as if source charset
- # is ISO8859-1 and not some other. This breaks our
- # isis character encoding, so we convert it first
- # back to ISO8859-1 (which can actually be different
- # encoding in isis)
-
- $swish_data .= $swish if ($s && $swish);
- $display_data .= $display if ($d && $display);
- $index_data .= $display if ($i && $display);
+ # what will separate last line from this one?
+ if ($display_data && $x->{append} && $x->{append} eq "1") {
+ $line_delimiter = ' ';
+ } elsif ($display_data) {
+ $line_delimiter = '
';
+ }
+
+ # init vars so that we go into while...
+ ($swish,$display) = (1,1);
+
+ while ($swish || $display) {
+ ($swish,$display) = parse_format($format,$row,$isis_i++,$isis2cp);
+
+ # filter="name" ; filter this field through
+ # filter/[name].pm
+ my $filter = $x->{filter};
+ if ($filter) {
+ require "filter/".$filter.".pm";
+ }
+ # type="swish" ; field for swish
+ if ($s && $swish) {
+ if ($filter) {
+ no strict 'refs';
+ $swish_data .= join(" ",&$filter($swish));
+print STDERR "#### $swish_data\n";
+ } else {
+ $swish_data .= $swish;
+ }
+ }
+
+ # type="display" ; field for display
+ if ($d && $display) {
+ if ($line_delimiter && $display_data) {
+ $display_data .= $line_delimiter;
+ undef $line_delimiter;
+ }
+ if ($filter) {
+ no strict 'refs';
+ $display_data .= join($delimiter,&$filter($display));
+ } else {
+ if ($display_data) {
+ $display_data .= $delimiter.$display;
+ } else {
+ $display_data .= $display;
+ }
+ }
+ }
+
+ # type="index" ; insert into index
+ if ($i && $display) {
+ my $index_data = $display;
+ if ($filter) {
+ no strict 'refs';
+ foreach my $d (&$filter($index_data)) {
+ $index->insert($field, $d, $db_dir);
+ }
+ } else {
+ $index->insert($field, $index_data, $db_dir);
+ }
+ }
+ }
}
-#print STDERR "s_d: $swish_data\nd_d: $display_data\n" if ($swish_data);
if ($display_data) {
- $display_data = $isis_codepage->convert($display_data) || die "Can't convert '$display_data' !";
- # FIX: this is removed and replaced by html tag.
- #$xml .= xmlify($field."_display", $display_data);
if ($field eq "headline") {
$xml .= xmlify("headline", $display_data);
@@ -113,29 +176,26 @@
$field_name = $config->{indexer}->{$field}->{name_singular}."#-#";
} elsif ($config->{indexer}->{$field}->{name_plural}) {
$field_name = $config->{indexer}->{$field}->{name_plural}."#-#";
- } else {
+ } elsif ($config->{indexer}->{$field}->{name}) {
$field_name = $config->{indexer}->{$field}->{name}."#-#";
+ } else {
+ print STDERR "WARNING: field '$field' doesn't have 'name' attribute!";
}
if ($field_name) {
- $html .= $xml_codepage->convert($field_name);
+ $html .= x($field_name);
}
$html .= $display_data."###\n";
}
}
if ($swish_data) {
- my $i = Text::Iconv->new($config->{isis_codepage},'ISO8859-2');
- $swish_data = $i->convert($swish_data);
- $xml .= xmlify($field."_swish",unac_string('ISO8859-2',$swish_data));
- #$swish_data = $isis_codepage->convert($swish_data)."##" || $swish_data;
- #$xml .= xmlify($field."_swish",unac_string($config->{isis_codepage},$swish_data));
- }
+ # remove extra spaces
+ $swish_data =~ s/ +/ /g;
+ $swish_data =~ s/ +$//g;
- # index
- if ($index_data && $index_data ne "") {
- $index_data = $index_codepage->convert($index_data) || $index_data;
- $index->insert($field, $index_data, $db_dir);
+ $xml .= xmlify($field."_swish", unac_string($codepage,$swish_data));
}
+
}
# dump formatted output in
@@ -144,7 +204,6 @@
}
if ($xml) {
-#print STDERR "x: $xml\n";
$xml .= $add_xml if ($add_xml);
return "\n$xml\n";
} else {
@@ -158,15 +217,13 @@
foreach my $database ($cfg->Sections) {
- my $isis_db = $cfg -> val($database, 'isis_db');
- my $type = $cfg -> val($database, 'type');
- my $add_xml = $cfg -> val($database, 'xml');
+ my $isis_db = $cfg -> val($database, 'isis_db') || die "$database doesn't have 'isis_db' defined!";
+ my $type = $cfg -> val($database, 'type') || die "$database doesn't have 'type' defined";
+ my $add_xml = $cfg -> val($database, 'xml'); # optional
- # read configuration for this type
$config=XMLin("./import_xml/$type.xml", forcearray => [ 'isis' ], forcecontent => 1);
- $isis_codepage = Text::Iconv->new($config->{isis_codepage},'UTF8');
- $index_codepage = Text::Iconv->new($config->{isis_codepage},$config->{index_codepage});
- $xml_codepage = Text::Iconv->new($cfg->val($database,'xml_codepage'),'UTF8');
+
+ $isis2cp = Text::Iconv->new($config->{isis_codepage},$codepage);
my $db = OpenIsis::open( $isis_db );
if (0) {
@@ -189,7 +246,6 @@
for (my $row_id = 1; $row_id <= $max_rowid; $row_id++ ) {
my $row = OpenIsis::read( $db, $row_id );
if ($row && $row->{mfn}) {
-#print STDERR "mfn: ",$row->{mfn},"\n";
# output current process indicator
my $p = int($row->{mfn} * 100 / $max_rowid);
if ($p != $last_p) {
@@ -197,8 +253,8 @@
$last_p = $p;
}
- if (my $xml = isis2xml($row,$add_xml)) {
-#print STDERR "--ret-->$xml\n";
+ if (my $xml = $cp2utf->convert(isis2xml($row,$add_xml))) {
+ use bytes; # as opposed to chars
print "Path-Name: $path#".int($row->{mfn})."\n";
print "Content-Length: ".(length($xml)+1)."\n";
print "Document-Type: XML\n\n$xml\n";