--- trunk/all2xml.pl 2003/03/15 21:33:36 40
+++ trunk/all2xml.pl 2004/09/29 17:22:24 488
@@ -5,22 +5,25 @@
use Getopt::Std;
use Data::Dumper;
use XML::Simple;
-use Text::Unaccent 1.02; # 1.01 won't compile on my platform,
use Text::Iconv;
use Config::IniFiles;
use Encode;
+#use GDBM_File;
+use Fcntl; # for O_RDWR
+use TDB_File;
$|=1;
my $config_file = $0;
$config_file =~ s/\.pl$/.conf/;
+$config_file = $ARGV[0] if ($ARGV[0] && -f $ARGV[0]);
die "FATAL: can't find configuration file '$config_file'" if (! -e $config_file);
my $config;
-use index_DBI; # there is no other, right now ;-)
-
-my $index = new index_DBI(); # open index
+#use index_DBI; # default DBI module for index
+use index_DBI_cache; # faster DBI module using memory cache
+my $index;
my %opts;
@@ -32,9 +35,9 @@
getopts('d:m:qs', \%opts);
-my $db_dir;
+my $path; # this is name of database
-Text::Iconv->raise_error(1); # Conversion errors raise exceptions
+Text::Iconv->raise_error(0); # Conversion errors don't raise exceptions
# 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!
@@ -46,18 +49,41 @@
return $utf2cp->convert($_[0]);
}
-# decode isis import codepage
-my $isis2cp;
+# decode isis/excel or other import codepage
+my $import2cp;
# outgoing xml must be in UTF-8
my $cp2utf = Text::Iconv->new($codepage,'UTF-8');
-sub isis2xml {
+# mapping between data type and tag which specify
+# format in XML file
+my %type2tag = (
+ 'isis' => 'isis',
+ 'excel' => 'column',
+ 'marc' => 'marc',
+ 'feed' => 'feed'
+);
+
+my $cache; # for cacheing
+
+# lookup hash (tied to file)
+my %lhash;
+# this option will cache all lookup entries in memory.
+# if you are tight on memory, turn this off
+my $use_lhash_cache = 1;
+
+my $last_field_name; # cache to prevent repeated fields
+
+sub data2xml {
use xmlify;
+ my $type = shift @_;
my $row = shift @_;
my $add_xml = shift @_;
+ # needed to read values from configuration file
+ my $cfg = shift @_;
+ my $database = shift @_;
my $xml;
@@ -69,40 +95,159 @@
# sort subrouting using order="" attribute
sub by_order {
- return 0 if (! $config->{indexer}->{$a}->{order});
- return 0 if (! $config->{indexer}->{$b}->{order});
+ my $va = $config->{indexer}->{$a}->{order} ||
+ $config->{indexer}->{$a};
+ my $vb = $config->{indexer}->{$b}->{order} ||
+ $config->{indexer}->{$b};
- return $config->{indexer}->{$a}->{order} <=>
- $config->{indexer}->{$b}->{order} ;
+ return $va <=> $vb;
}
- foreach my $field (sort by_order keys %{$config->{indexer}}) {
+ my @sorted_tags;
+ if ($cache->{tags_by_order}) {
+ @sorted_tags = @{$cache->{tags_by_order}};
+ } else {
+ @sorted_tags = sort by_order keys %{$config->{indexer}};
+ $cache->{tags_by_order} = \@sorted_tags;
+ }
+
+ # lookup key
+ my $lookup_key;
+
+ # cache for field in pages
+ delete $cache->{display_data};
+ delete $cache->{swish_data};
+ delete $cache->{swish_exact_data};
+ delete $cache->{index_data};
+ delete $cache->{index_delimiter};
+ my @page_fields; # names of fields
+
+
+ # subs used to produce output
+
+ sub get_field_name($$$) {
+ my ($config,$field,$field_usage) = @_;
+
+ # find field name (signular, plural)
+ my $field_name = "";
+ if ($config->{indexer}->{$field}->{name_singular} && $field_usage == 1) {
+ $field_name = $config->{indexer}->{$field}->{name_singular};
+ } elsif ($config->{indexer}->{$field}->{name_plural}) {
+ $field_name = $config->{indexer}->{$field}->{name_plural};
+ } 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) {
+ if (! $last_field_name) {
+ $last_field_name = x($field_name);
+ return $last_field_name;
+ } elsif ($field_name ne $last_field_name) {
+ $last_field_name = x($field_name);
+ return $last_field_name;
+ }
+ }
+ }
- $field=x($field);
+ # init variables for different types
+ sub init_visible_type($) {
+ my $type = shift;
+
+ # swish, swish_exact, display, index, index_lookup
+ # swish and display defaults
+ my ($s,$se,$d,$i,$il) = (1,0,1,0,0);
+ if (lc($type) eq "display") {
+ $s = 0;
+ } elsif (lc($type) eq "swish") {
+ $d = 0;
+ } elsif (lc($type) eq "index") {
+ ($s,$se,$d,$i) = (0,1,0,1);
+ } elsif (lc($type) eq "swish_exact") {
+ ($s,$se,$d,$i) = (0,1,0,0);
+ } elsif (lc($type) =~ /^lookup/) {
+ ($s,$se,$d,$i,$il) = (0,1,0,0,1);
+ } elsif ($type) {
+ print STDERR "WARNING: unknown type: $type\n";
+ }
+ return ($s,$se,$d,$i,$il);
+ }
+
+
+ # convert
+ #
+ #
+ # ,
+ # 200a
+ #
+ #
+ # to
+ #
+ # 200a
+ #
+ # but without loosing spaces in delimiter (becasue
+ # new XML::Simple strips spaces in attribute values
+ # as defined in XML specification)
+ #
+ sub unroll_x($) {
+ my $x = shift;
+
+ if (defined $x->{value}) {
+ my ($v,$d) = ($x->{value}->{content}, $x->{delimiter}->{content});
+ delete $x->{value};
+ delete $x->{delimiter};
+ $x->{content} = $v;
+ $x->{delimiter} = $d;
+ }
+ return $x;
+ }
+
+ # begin real work: go field by field
+ foreach my $field (@sorted_tags) {
+
+ $field=x($field);
$field_usage{$field}++;
my $swish_data = "";
+ my $swish_exact_data = "";
my $display_data = "";
+ my @index_data;
my $line_delimiter;
my ($swish,$display);
- foreach my $x (@{$config->{indexer}->{$field}->{isis}}) {
+ my $tag = $type2tag{$type} || die "can't find which tag to use for type $type";
+
+ # is this field page-by-page?
+ my $iterate_by_page = $config->{indexer}->{$field}->{iterate_by_page};
+ push @page_fields,$field if ($iterate_by_page);
+ my %page_max = ();
+ # default line_delimiter if using
+ my $page_line_delimiter = $config->{indexer}->{$field}->{page_line_delimiter} || '
';
+ $cache->{index_delimiter}->{$field} = $config->{indexer}->{$field}->{index_delimiter};
+
+ my $format_name = $config->{indexer}->{$field}->{format_name};
+ my $format_delimiter = $config->{indexer}->{$field}->{format_delimiter};
+ if ($format_name && $format_delimiter) {
+ $cache->{format}->{$field}->{format_name} = $format_name;
+ $cache->{format}->{$field}->{format_delimiter} = $format_delimiter;
+ }
+
+ foreach my $x (@{$config->{indexer}->{$field}->{$tag}}) {
+
+ $x = unroll_x($x);
my $format = x($x->{content});
my $delimiter = x($x->{delimiter}) || ' ';
- my $isis_i = 0; # isis repeatable offset
+ my $repeat_off = 0; # init 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");
+ my ($s,$se,$d,$i,$il) = init_visible_type($x->{type});
# what will separate last line from this one?
- if ($display_data && $x->{append} && $x->{append} eq "1") {
- $line_delimiter = ' ';
+ if ($display_data && $x->{append}) {
+ $line_delimiter = $delimiter;
} elsif ($display_data) {
$line_delimiter = '
';
}
@@ -110,96 +255,339 @@
# init vars so that we go into while...
($swish,$display) = (1,1);
+ # placeholder for all repeatable entries for index
+
+ sub mkformat($$) {
+ my $x = shift || die "mkformat needs tag reference";
+ my $data = shift || return;
+ my $format_name = x($x->{format_name}) || return $data;
+ my $fmt = x($config->{format}->{$format_name}->{content}) || die " is not defined!";
+ my $format_delimiter = x($x->{format_delimiter});
+ my @data;
+ if ($format_delimiter) {
+ @data = split(/$format_delimiter/,$data);
+ } else {
+ push @data,$data;
+ }
+
+ if ($fmt) {
+ my $nr = scalar $fmt =~ s/%s/%s/g;
+ if (($#data+1) == $nr) {
+ return sprintf($fmt,@data);
+ } else {
+ #print STDERR "mkformat: [$data] can't be split on [$format_delimiter] to $nr fields!\n";
+ return $data;
+ }
+ } else {
+ print STDERR "usage of link '$format_name' without defined format ( tag)\n";
+ }
+ }
+
+ # while because of repeatable fields
while ($swish || $display) {
- ($swish,$display) = parse_format($format,$row,$isis_i++,$isis2cp);
+ my $page = $repeat_off;
+ $page_max{$field} = $page if ($iterate_by_page && $page > ($page_max{$field} || 0));
+ ($swish,$display) = parse_format($type, $format,$row,$repeat_off++,$import2cp);
+ if ($repeat_off > 1000) {
+ print STDERR "loop (more than 1000 repeatable fields) deteced in $row, $format\n";
+ last;
+ }
+
+ # is this field is lookup?
+ if ($display && $x->{lookup}) {
+ my $null = "";
+ if ($use_lhash_cache) {
+ if (!defined($cache->{lhash}->{$display})) {
+ my $new_display = $lhash{$display};
+ if (defined($new_display)) {
+#print STDERR "lookup cache store '$display' = '$new_display'\n";
+ $display = $new_display;
+ $cache->{lhash}->{$display} = $new_display;
+ } else {
+# print STDERR "WARNING: lookup for '$display' didn't find anything.\n";
+ $display = "";
+ $cache->{lhash}->{$display} = $null;
+ }
+ } else {
+ $display = $cache->{lhash}->{$display};
+ }
+ } else {
+ $display = $lhash{$display} || $null;
+ }
+ }
# filter="name" ; filter this field through
# filter/[name].pm
my $filter = $x->{filter};
- if ($filter) {
+ if ($filter && !$cache->{filter_loaded}->{$filter}) {
require "filter/".$filter.".pm";
+ $cache->{filter_loaded}->{$filter}++;
}
# type="swish" ; field for swish
- if ($s && $swish) {
- if ($filter) {
+ if ($swish) {
+ my $tmp = $swish;
+ if ($filter && ($s || $se)) {
no strict 'refs';
- $swish_data .= join(" ",&$filter($swish));
- } else {
- $swish_data .= $swish;
+ $tmp = join(" ",&$filter($tmp)) if ($s || $se);
}
+
+ $swish_data .= $tmp if ($s && $tmp);
+ $swish_exact_data .= "xxbxx $tmp xxexx " if ($tmp && $tmp ne "" && $se);
}
# type="display" ; field for display
if ($d && $display) {
+ my $ldel = $delimiter;
if ($line_delimiter && $display_data) {
- $display_data .= $line_delimiter;
- undef $line_delimiter;
+ $ldel = $line_delimiter;
}
if ($filter) {
no strict 'refs';
- $display_data .= join($delimiter,&$filter($display));
- } else {
- if ($display_data) {
- $display_data .= $delimiter.$display;
- } else {
- $display_data .= $display;
+ my @arr;
+ foreach my $tmp (&$filter($display)) {
+ my $tmp2 = mkformat($x,$tmp);
+ push @arr,$tmp2 if ($tmp2);
}
+ $display_data .= $ldel if ($display_data && @arr);
+ $display_data .= join($delimiter,@arr);
+ } else {
+ $display_data .= $ldel if ($display_data);
+ my $tmp = mkformat($x,$display);
+ $display_data .= $tmp if ($tmp);
}
}
# type="index" ; insert into index
+ my $idisplay;
if ($i && $display) {
- my $index_data = $display;
+ $idisplay = $display;
if ($filter) {
no strict 'refs';
- foreach my $d (&$filter($index_data)) {
- $index->insert($field, $d, $db_dir);
+ $idisplay = &$filter($idisplay);
+ }
+ push @index_data, $idisplay if ($idisplay && !$iterate_by_page);
+ }
+
+ # store fields in lookup
+ if ($il && $display) {
+ if (lc($x->{type}) eq "lookup_key") {
+ if ($lookup_key) {
+ print STDERR "WARNING: try to redefine lookup_key (keys shouldn't be repeatable fields!)";
+ } else {
+ if ($filter) {
+ no strict 'refs';
+ $lookup_key = &$filter($display);
+ } else {
+ $lookup_key = $display;
+ }
+ }
+ } elsif (lc($x->{type}) eq "lookup_val") {
+ if ($lookup_key) {
+ if ($filter) {
+ no strict 'refs';
+ $lhash{$lookup_key} = &$filter($display);
+ } else {
+ $lhash{$lookup_key} = $display;
+ }
+ } else {
+ print STDERR "WARNING: no lookup_key defined for '$display'?";
}
- } else {
- $index->insert($field, $index_data, $db_dir);
}
+
}
+
+ # store data for page-by-page repeatable fields
+ if ($iterate_by_page) {
+ sub iterate_fld($$$$$$) {
+ my ($cache,$what,$field,$page,$data,$append) = @_;
+ return if (!$data);
+
+ my $ldel = $page_line_delimiter;
+ $ldel = " " if ($append);
+#print STDERR "line delimiter: ",Dumper($ldel) if ($ldel);
+ if (! $cache->{$what}->{$field}->[$page]) {
+ $cache->{$what}->{$field}->[$page] = $data;
+ } else {
+ $cache->{$what}->{$field}->[$page] .= $ldel.$data;
+ }
+ }
+
+ if ($display_data) {
+ iterate_fld($cache,'display_data',$field,$page,$display_data,$x->{append});
+ }
+ $display_data = "";
+ if ($swish_data) {
+ iterate_fld($cache,'swish_data',$field,$page,$swish_data,$x->{append});
+ $swish_data = "";
+ }
+ if ($swish_exact_data) {
+ iterate_fld($cache,'swish_exact_data',$field,$page,$swish_exact_data,$x->{append});
+ $swish_exact_data = "";
+ }
+
+ if ($idisplay) {
+ my $ldel=$page_line_delimiter;
+ my @index_data;
+ if ($cache->{index_data}->{$field}->[$page]) {
+
+ @index_data = @{$cache->{index_data}->{$field}->[$page]};
+ }
+ if ($x->{append}) {
+ if (@index_data) {
+ $index_data[$#index_data] .= $idisplay;
+ } else {
+ push @index_data, $idisplay;
+ }
+ } else {
+ push @index_data, $idisplay;
+ }
+ $idisplay = "";
+ @{$cache->{index_data}->{$field}->[$page]} = @index_data;
+ }
+ }
+ }
+
+ if (! $iterate_by_page) {
+ my $idel = $x->{index_delimiter};
+ # fill data in index
+ foreach my $tmp (@index_data) {
+ my $i = $d = $tmp;
+ if ($idel && $tmp =~ m/$idel/) {
+ ($i,$d) = split(/$idel/,$tmp);
+ }
+ $index->insert($field, $i, $d, $path);
+ }
+ @index_data = ();
}
}
+ # now try to parse variables from configuration file
+ foreach my $x (@{$config->{indexer}->{$field}->{'config'}}) {
- if ($display_data) {
+ $x = unroll_x($x);
- if ($field eq "headline") {
- $xml .= xmlify("headline", $display_data);
- } else {
+ my $delimiter = x($x->{delimiter}) || ' ';
+ my $val = $cfg->val($database, x($x->{content}));
- # find field name (signular, plural)
- my $field_name = "";
- if ($config->{indexer}->{$field}->{name_singular} && $field_usage{$field} == 1) {
- $field_name = $config->{indexer}->{$field}->{name_singular}."#-#";
- } elsif ($config->{indexer}->{$field}->{name_plural}) {
- $field_name = $config->{indexer}->{$field}->{name_plural}."#-#";
- } elsif ($config->{indexer}->{$field}->{name}) {
- $field_name = $config->{indexer}->{$field}->{name}."#-#";
- } else {
- print STDERR "WARNING: field '$field' doesn't have 'name' attribute!";
+ # FIXME index_lookup is not supported!
+ my ($s,$se,$d,$i,$il) = init_visible_type($x->{type});
+
+ if ($val) {
+ $display_data .= $delimiter.$val if ($d);
+ $swish_data .= " ".$val if ($s);
+ $index->insert($field, $val, $path) if ($i);
+ }
+
+ if ($iterate_by_page) {
+ # FIXME data from config tag will appear just
+ # on first page!!!
+ my $page = 0;
+ if ($display_data) {
+ $cache->{display_data}->{$field}->[$page] = $display_data;
+ $display_data = "";
}
- if ($field_name) {
- $html .= x($field_name);
+ if ($swish_data) {
+ $cache->{swish_data}->{$field}->[$page] = $swish_data;
+ $swish_data = "";
+ }
+ if ($swish_exact_data) {
+ $cache->{swish_exact_data}->{$field}->[$page] = $swish_exact_data;
+ $swish_exact_data = "";
}
- $html .= $display_data."###\n";
}
}
- if ($swish_data) {
- # remove extra spaces
- $swish_data =~ s/ +/ /g;
- $swish_data =~ s/ +$//g;
- $xml .= xmlify($field."_swish", unac_string($codepage,$swish_data));
+ # save data page-by-page
+ foreach my $field (@page_fields) {
+ my $nr_pages = $page_max{$field} || next;
+#print STDERR "field '$field' iterate over ",($nr_pages || 0)," pages...\n";
+#print STDERR Dumper($cache->{display_data});
+ for (my $page=0; $page <= $nr_pages; $page++) {
+ my $display_data;
+ if ($cache->{format}->{$field}) {
+ my $tmp = mkformat($cache->{format}->{$field},$cache->{display_data}->{$field}->[$page]);
+ $display_data=$tmp if ($tmp);
+ } else {
+ $display_data = $cache->{display_data}->{$field}->[$page];
+ }
+ if ($display_data) { # default
+ if ($field eq "headline") {
+ $xml .= xmlify("headline", $display_data);
+ } else {
+
+ # fallback to empty field name if needed
+ $html .= get_field_name($config,$field,$field_usage{$field}) || '';
+ $html .= "#-#".$display_data."###\n";
+ }
+ }
+
+ my $swish_data = $cache->{swish_data}->{$field}->[$page];
+ if ($swish_data) {
+ # remove extra spaces
+ $swish_data =~ s/ +/ /g;
+ $swish_data =~ s/ +$//g;
+
+ $xml .= xmlify($field."_swish", my_unac_string($codepage,$swish_data));
+ }
+
+ my $swish_exact_data = $cache->{swish_exact_data}->{$field}->[$page];
+ if ($swish_exact_data) {
+ $swish_exact_data =~ s/ +/ /g;
+ $swish_exact_data =~ s/ +$//g;
+
+ # add delimiters before and after word.
+ # That is required to produce exact match
+ $xml .= xmlify($field."_swish_exact", my_unac_string($codepage,$swish_exact_data));
+ }
+
+ my $idel = $cache->{index_delimiter}->{$field};
+ foreach my $tmp (@{$cache->{index_data}->{$field}->[$page]}) {
+ my $i = $tmp;
+ my $d = $tmp;
+ if ($idel && $tmp =~ m/$idel/) {
+ ($i,$d) = split(/$idel/,$tmp);
+ }
+ $index->insert($field, $i, $d, $path);
+#print STDERR "index [$idel] $field: $i --> $d [$path]\n";
+ }
+ }
+
}
+
+ if (! $iterate_by_page) {
+ if ($display_data) {
+ if ($field eq "headline") {
+ $xml .= xmlify("headline", $display_data);
+ } else {
+
+ # fallback to empty field name if needed
+ $html .= get_field_name($config,$field,$field_usage{$field}) || '';
+ $html .= "#-#".$display_data."###\n";
+ }
+ }
+ if ($swish_data) {
+ # remove extra spaces
+ $swish_data =~ s/ +/ /g;
+ $swish_data =~ s/ +$//g;
+ $xml .= xmlify($field."_swish", my_unac_string($codepage,$swish_data));
+ }
+ if ($swish_exact_data) {
+ $swish_exact_data =~ s/ +/ /g;
+ $swish_exact_data =~ s/ +$//g;
+
+ # add delimiters before and after word.
+ # That is required to produce exact match
+ $xml .= xmlify($field."_swish_exact", my_unac_string($codepage,$swish_exact_data));
+ }
+ }
}
# dump formatted output in
if ($html) {
- $xml .= xmlify("html",$html);
+ #$xml .= xmlify("html",$html);
+ $xml .= "";
}
if ($xml) {
@@ -212,55 +600,330 @@
##########################################################################
+# read configuration for this script
my $cfg = new Config::IniFiles( -file => $config_file );
+# read global.conf configuration
+my $cfg_global = new Config::IniFiles( -file => 'global.conf' );
+
+# open index
+$index = new index_DBI(
+ $cfg_global->val('global', 'dbi_dbd'),
+ $cfg_global->val('global', 'dbi_dsn'),
+ $cfg_global->val('global', 'dbi_user'),
+ $cfg_global->val('global', 'dbi_passwd') || '',
+ );
+
+my $show_progress = $cfg_global->val('global', 'show_progress');
+
+my $my_unac_filter = $cfg_global->val('global', 'my_unac_filter');
+if ($my_unac_filter) {
+ print STDERR "using $my_unac_filter to filter characters for search\n";
+ require $my_unac_filter;
+} else {
+ print STDERR "### fallback to default my_unac_string!\n";
+ eval q{
+ sub main::my_unac_string($$) {
+ my ($charset, $string) = (@_);
+ return $string;
+ }
+ };
+}
+
foreach my $database ($cfg->Sections) {
- 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 $type = lc($cfg -> val($database, 'type')) || die "$database doesn't have 'type' defined";
my $add_xml = $cfg -> val($database, 'xml'); # optional
- $config=XMLin("./import_xml/$type.xml", forcearray => [ 'isis' ], forcecontent => 1);
-
- $isis2cp = Text::Iconv->new($config->{isis_codepage},$codepage);
+ # create new lookup file
+ my $lookup_file = $cfg -> val($database, 'lookup_newfile'); # optional
+ if ($lookup_file) {
+ #tie %lhash, 'GDBM_File', $lookup_file, &GDBM_NEWDB, 0644;
+ if (! -e $lookup_file) {
+ open(LOOKUP, "> $lookup_file") || die "can't create $lookup_file': $!";
+ close(LOOKUP);
+ }
+ tie %lhash, 'TDB_File', $lookup_file, TDB_CLEAR_IF_FIRST, O_RDWR, 0644;
+ print STDERR "creating lookup file '$lookup_file'\n";
+ # delete memory cache for lookup file
+ delete $cache->{lhash};
+ }
- my $db = OpenIsis::open( $isis_db );
- if (0) {
-# # FIX
-# if (! $db ) {
- print STDERR "WARNING: can't open '$isis_db'\n";
- next ;
+ # open existing lookup file
+ $lookup_file = $cfg -> val($database, 'lookup_open'); # optional
+ if ($lookup_file) {
+ #tie %lhash, 'GDBM_File', $lookup_file, &GDBM_READER, 0644;
+ tie %lhash, 'TDB_File', $lookup_file, TDB_DEFAULT, O_RDWR, 0644;
+ print STDERR "opening lookup file '$lookup_file'\n";
}
- my $max_rowid = OpenIsis::maxRowid( $db );
+print STDERR "reading ./import_xml/$type.xml\n";
- print STDERR "Reading database: $isis_db [$max_rowid rows]\n";
+ # extract just type basic
+ my $type_base = $type;
+ $type_base =~ s/_.+$//g;
- my $path = $database; # was $isis_db
+ $config=XMLin("./import_xml/$type.xml", ForceArray => [ $type2tag{$type_base}, 'config', 'format' ], ForceContent => 1 );
+ # output current progress indicator
my $last_p = 0;
+ sub progress {
+ return if (! $show_progress);
+ my $current = shift;
+ my $total = shift || 1;
+ my $p = int($current * 100 / $total);
+ if ($p != $last_p) {
+ printf STDERR ("%5d / %5d [%-51s] %-2d %% \r",$current,$total,"=" x ($p/2).">", $p );
+ $last_p = $p;
+ }
+ }
+
+ my $fake_dir = 1;
+ sub fakeprogress {
+ return if (! $show_progress);
+ my $current = shift @_;
+
+ my @ind = ('-','\\','|','/','-','\\','|','/', '-');
+
+ $last_p += $fake_dir;
+ $fake_dir = -$fake_dir if ($last_p > 1000 || $last_p < 0);
+ if ($last_p % 10 == 0) {
+ printf STDERR ("%5d / %5s [%-51s]\r",$current,"?"," " x ($last_p/20).$ind[($last_p/20) % $#ind]);
+ }
+ }
-# { my $row_id = 4514;
-# FIX
- for (my $row_id = 1; $row_id <= $max_rowid; $row_id++ ) {
- my $row = OpenIsis::read( $db, $row_id );
- if ($row && $row->{mfn}) {
- # output current process indicator
- my $p = int($row->{mfn} * 100 / $max_rowid);
- if ($p != $last_p) {
- printf STDERR ("%5d / %5d [%-51s] %-2d %% \r",$row->{mfn},$max_rowid,"=" x ($p/2).">", $p ) if (! $opts{q});
- $last_p = $p;
+ # now read database
+print STDERR "using: $type...\n";
+
+ # erase cache for tags by order in this database
+ delete $cache->{tags_by_order};
+
+ if ($type_base eq "isis") {
+
+ my $isis_db = $cfg -> val($database, 'isis_db') || die "$database doesn't have 'isis_db' defined!";
+
+ $import2cp = Text::Iconv->new($config->{isis_codepage},$codepage);
+ my $db = OpenIsis::open( $isis_db );
+
+ # check if .txt database for OpenIsis is zero length,
+ # if so, erase it and re-open database
+ sub check_txt_db {
+ my $isis_db = shift || die "need isis database name";
+ my $reopen = 0;
+
+ if (-e $isis_db.".TXT") {
+ print STDERR "WARNING: removing $isis_db.TXT OpenIsis database...\n";
+ unlink $isis_db.".TXT" || warn "FATAL: unlink error on '$isis_db.TXT': $!";
+ $reopen++;
}
+ if (-e $isis_db.".PTR") {
+ print STDERR "WARNING: removing $isis_db.PTR OpenIsis database...\n";
+ unlink $isis_db.".PTR" || warn "FATAL: unlink error on '$isis_db.PTR': $!";
+ $reopen++;
+ }
+ return OpenIsis::open( $isis_db ) if ($reopen);
+ }
+
+ # EOF error
+ if ($db == -1) {
+ $db = check_txt_db($isis_db);
+ if ($db == -1) {
+ print STDERR "FATAL: OpenIsis can't open zero size file $isis_db\n";
+ next;
+ }
+ }
- if (my $xml = $cp2utf->convert(isis2xml($row,$add_xml))) {
+ # OpenIsis::ERR_BADF
+ if ($db == -4) {
+ print STDERR "FATAL: OpenIsis can't find file $isis_db\n";
+ next;
+ # OpenIsis::ERR_IO
+ } elsif ($db == -5) {
+ print STDERR "FATAL: OpenIsis can't access file $isis_db\n";
+ next;
+ } elsif ($db < 0) {
+ print STDERR "FATAL: OpenIsis unknown error $db with file $isis_db\n";
+ next;
+ }
+
+ my $max_rowid = OpenIsis::maxRowid( $db );
+
+ # if 0 records, try to rease isis .txt database
+ if ($max_rowid == 0) {
+ # force removal of database
+ $db = check_txt_db($isis_db);
+ $max_rowid = OpenIsis::maxRowid( $db );
+ }
+
+ print STDERR "Reading database: $isis_db [$max_rowid rows]\n";
+
+ my $path = $database;
+
+ for (my $row_id = 1; $row_id <= $max_rowid; $row_id++ ) {
+ my $row = OpenIsis::read( $db, $row_id );
+ if ($row && $row->{mfn}) {
+
+ progress($row->{mfn}, $max_rowid);
+
+ my $swishpath = $path."#".int($row->{mfn});
+
+ if (my $xml = data2xml($type_base,$row,$add_xml,$cfg,$database)) {
+ $xml = $cp2utf->convert($xml);
+ use bytes; # as opposed to chars
+ print "Path-Name: $swishpath\n";
+ print "Content-Length: ".(length($xml)+1)."\n";
+ print "Document-Type: XML\n\n$xml\n";
+ }
+ }
+ }
+ # for this to work with current version of OpenIsis (0.9.0)
+ # you might need my patch from
+ # http://www.rot13.org/~dpavlin/projects/openisis-0.9.0-perl_close.diff
+ OpenIsis::close($db);
+ print STDERR "\n";
+
+ } elsif ($type_base eq "excel") {
+ require Spreadsheet::ParseExcel;
+ require Spreadsheet::ParseExcel::Utility;
+ import Spreadsheet::ParseExcel::Utility qw(int2col);
+
+ $import2cp = Text::Iconv->new($config->{excel_codepage},$codepage);
+ my $excel_file = $cfg -> val($database, 'excel_file') || die "$database doesn't have 'excel_file' defined!";
+
+ my $sheet = x($config->{sheet}) || die "no sheet in $type.xml";
+ my $start_row = x($config->{start_row}) - 1 || die "no start_row in $type.xml";
+
+ my $oBook = Spreadsheet::ParseExcel::Workbook->Parse($excel_file) || die "can't open Excel file '$excel_file'";
+
+ my $sheet_nr = 0;
+ foreach my $oWks (@{$oBook->{Worksheet}}) {
+ #print STDERR "-- SHEET $sheet_nr:", $oWks->{Name}, "\n";
+ last if ($oWks->{Name} eq $sheet);
+ $sheet_nr++;
+ }
+
+ my $oWorksheet = $oBook->{Worksheet}[$sheet_nr];
+
+ print STDERR "using sheet: ",$oWorksheet->{Name},"\n";
+ defined ($oWorksheet) || die "can't find sheet '$sheet' in $excel_file";
+ my $end_row = x($config->{end_row}) || $oWorksheet->{MaxRow};
+
+ for(my $iR = $start_row ; defined $end_row && $iR <= $end_row ; $iR++) {
+ my $row;
+ for(my $iC = $oWorksheet->{MinCol} ; defined $oWorksheet->{MaxCol} && $iC <= $oWorksheet->{MaxCol} ; $iC++) {
+ my $cell = $oWorksheet->{Cells}[$iR][$iC];
+ if ($cell) {
+ # this conversion is a cludge.
+ # Files from Excell could have
+ # characters which don't fit into
+ # destination encoding.
+ $row->{int2col($iC)} = $utf2cp->convert($cell->Value) || $cell->Value;
+ }
+ }
+
+ progress($iR, $end_row);
+
+# print "row[$iR/$end_row] ";
+# foreach (keys %{$row}) {
+# print "$_: ",$row->{$_},"\t";
+# }
+# print "\n";
+
+ my $swishpath = $database."#".$iR;
+
+ next if (! $row);
+
+ if (my $xml = data2xml($type_base,$row,$add_xml,$cfg,$database)) {
+ $xml = $cp2utf->convert($xml);
use bytes; # as opposed to chars
- print "Path-Name: $path#".int($row->{mfn})."\n";
+ print "Path-Name: $swishpath\n";
print "Content-Length: ".(length($xml)+1)."\n";
print "Document-Type: XML\n\n$xml\n";
}
}
+ } elsif ($type_base eq "marc") {
+
+ require MARC;
+
+ $import2cp = Text::Iconv->new($config->{marc_codepage},$codepage);
+ my $marc_file = $cfg -> val($database, 'marc_file') || die "$database doesn't have 'marc_file' defined!";
+
+ # optional argument is format
+ my $format = x($config->{marc_format}) || 'usmarc';
+
+ print STDERR "Reading MARC file '$marc_file'\n";
+
+ my $marc = new MARC;
+ my $nr = $marc->openmarc({
+ file=>$marc_file, format=>$format
+ }) || die "Can't open MARC file '$marc_file' with format '$format'";
+
+ # read MARC file in memory
+ $marc->nextmarc(-1);
+
+ my $max_rec = $marc->marc_count();
+
+ for(my $i=1; $i<=$max_rec; $i++) {
+
+ progress($i,$max_rec);
+
+ # store value for marc_sf.pm
+ $main::cache->{marc_record} = $i;
+
+ my $swishpath = $database."#".$i;
+
+ if (my $xml = data2xml($type_base,$marc,$add_xml,$cfg,$database)) {
+ $xml = $cp2utf->convert($xml);
+ use bytes; # as opposed to chars
+ print "Path-Name: $swishpath\n";
+ print "Content-Length: ".(length($xml)+1)."\n";
+ print "Document-Type: XML\n\n$xml\n";
+ }
+ }
+
+ print STDERR "\n";
+
+ } elsif ($type_base eq "feed") {
+
+ $import2cp = Text::Iconv->new($config->{feed_codepage},$codepage);
+ my $prog = x($config->{prog}) || die "$database doesn't have 'prog' defined!";
+
+ print STDERR "Reading feed from program '$prog'\n";
+
+ open(FEED,"feeds/$prog |") || die "can't start $prog: $!";
+
+ my $i=1; # record nr.
+
+ my $data;
+ my $line=1;
+
+ while () {
+ chomp;
+
+ if (/^$/) {
+ my $swishpath = $database."#".$i++;
+
+ if (my $xml = data2xml($type_base,$data,$add_xml,$cfg,$database)) {
+ $xml = $cp2utf->convert($xml);
+ use bytes; # as opposed to chars
+ print "Path-Name: $swishpath\n";
+ print "Content-Length: ".(length($xml)+1)."\n";
+ print "Document-Type: XML\n\n$xml\n";
+ }
+ $line = 1;
+ $data = {};
+ next;
+ }
+
+ $line = $1 if (s/^(\d+):\s*//);
+ $data->{$line++} = $_;
+
+ fakeprogress($i);
+
+ }
+ # close lookup
+ untie %lhash if (%lhash);
}
- print STDERR "\n";
}
# call this to commit index
@@ -272,15 +935,28 @@
=head1 NAME
-isis2xml.pl - read isis file and dump XML
+all2xml.pl - read various file formats and dump XML for SWISH-E
+
+=head1 SYNOPSYS
+
+ $ all2xml.pl [test.conf]
=head1 DESCRIPTION
-This command will read ISIS data file using OpenIsis perl module and
-create XML file for usage with I
-indexer. Dispite it's name, this script B
-from isis files (isis allready has something like that). Output of this
-script is tailor-made for SWISH-E.
+This command will read ISIS data file using OpenIsis perl module, MARC
+records using MARC module and optionally Micro$oft Excel files to
+create one XML file for usage with I indexer. Dispite it's name,
+this script B from isis files (isis allready
+has something like that). Output of this script is tailor-made for SWISH-E.
+
+If no configuration file is specified, it will use default one called
+C.
+
+=head1 BUGS
+
+Documentation is really lacking. However, in true Open Source spirit, source
+is best documentation. I even made considerable effort to comment parts
+which are not intuitively clear, so...
=head1 AUTHOR