#!/usr/bin/perl -w
use strict;
-use OpenIsis;
+use IsisDB;
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;
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;
# 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;
delete $cache->{display_data};
delete $cache->{swish_data};
delete $cache->{swish_exact_data};
+ delete $cache->{index_data};
+ delete $cache->{index_delimiter};
my @page_fields; # names of fields
} else {
print STDERR "WARNING: field '$field' doesn't have 'name' attribute!";
}
+
if ($field_name) {
- return x($field_name);
+ $field_name = x($field_name);
+ if (! $last_field_name) {
+ $last_field_name = $field_name;
+ return $last_field_name;
+ } elsif ($field_name ne $last_field_name) {
+ $last_field_name = $field_name;
+ return $last_field_name;
+ }
+ }
+ }
+
+
+ # init variables for different types
+ sub init_visible_type($) {
+ 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
+ #
+ # <tag>
+ # <delimiter>, </delimiter>
+ # <value>200a</value>
+ # </tag>
+ #
+ # to
+ #
+ # <tag delimiter=", ">200a</tag>
+ #
+ # 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) {
my $swish_data = "";
my $swish_exact_data = "";
my $display_data = "";
+ my @index_data;
my $line_delimiter;
my ($swish,$display);
my %page_max = ();
# default line_delimiter if using
my $page_line_delimiter = $config->{indexer}->{$field}->{page_line_delimiter} || '<br/>';
+ $cache->{index_delimiter}->{$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 $repeat_off = 0; # init repeatable offset
- # swish, swish_exact, display, index, index_lookup
- # swish and display defaults
- my ($s,$se,$d,$i,$il) = (1,0,1,0,0);
- $s = 0 if (lc($x->{type}) eq "display");
- $d = 0 if (lc($x->{type}) eq "swish");
- $se = 1 if (lc($x->{type}) eq "swish_exact");
- ($s,$d,$i) = (0,0,1) if (lc($x->{type}) eq "index");
- $il = 1 if (lc($x->{type}) =~ /^lookup/);
-
+ 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) {
$line_delimiter = '<br/>';
}
($swish,$display) = (1,1);
# placeholder for all repeatable entries for index
- my @index_data;
- sub mkformat {
+ sub mkformat($$) {
my $x = shift || die "mkformat needs tag reference";
my $data = shift || return;
my $format_name = x($x->{format_name}) || return $data;
if (($#data+1) == $nr) {
return sprintf($fmt,@data);
} else {
- print STDERR "mkformat: [$data] can't be split on [$format_delimiter] to $nr fields!\n";
+ #print STDERR "mkformat: [$data] can't be split on [$format_delimiter] to $nr fields!\n";
return $data;
}
} else {
$display = $new_display;
$cache->{lhash}->{$display} = $new_display;
} else {
- print STDERR "WARNING: lookup for '$display' didn't find anything.\n";
+# print STDERR "WARNING: lookup for '$display' didn't find anything.\n";
$display = "";
$cache->{lhash}->{$display} = $null;
}
}
# type="swish" ; field for swish
if ($swish) {
+ my $tmp = $swish;
if ($filter && ($s || $se)) {
no strict 'refs';
- my $tmp = join(" ",&$filter($swish)) if ($s || $se);
- $swish_data .= $tmp if ($s);
- $swish_exact_data .= $tmp if ($se);
-
- } else {
- $swish_data .= $swish if ($s);
- $swish_exact_data .= $swish if ($se);
+ $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;
+ $ldel = $line_delimiter;
}
if ($filter) {
no strict 'refs';
- if ($display_data) {
- $display_data .= $delimiter.join($delimiter,mkformat($x,&$filter($display)));
- } else {
- $display_data = join($delimiter,mkformat($x,&$filter($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 {
- if ($display_data) {
- $display_data .= $delimiter.mkformat($x,$display);
- } else {
- $display_data = mkformat($x,$display);
- }
+ $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) {
+ $idisplay = $display;
if ($filter) {
no strict 'refs';
- $display = &$filter($display);
- }
- if ($x->{append} && @index_data) {
- $index_data[$#index_data].=$display;
- } else {
- push @index_data, $display;
+ $idisplay = &$filter($idisplay);
}
+ push @index_data, $idisplay if ($idisplay && !$iterate_by_page);
}
# store fields in lookup
if ($lookup_key) {
print STDERR "WARNING: try to redefine lookup_key (keys shouldn't be repeatable fields!)";
} else {
- $lookup_key = $display;
+ if ($filter) {
+ no strict 'refs';
+ $lookup_key = &$filter($display);
+ } else {
+ $lookup_key = $display;
+ }
}
} elsif (lc($x->{type}) eq "lookup_val") {
if ($lookup_key) {
- $lhash{$lookup_key} = $display;
+ 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'?";
}
sub iterate_fld($$$$$$) {
my ($cache,$what,$field,$page,$data,$append) = @_;
return if (!$data);
- my $line_delimiter = $page_line_delimiter;
- $line_delimiter = " " if ($append);
+
+ 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] .= $line_delimiter.$data;
+ $cache->{$what}->{$field}->[$page] .= $ldel.$data;
}
}
if ($display_data) {
-print STDERR "line delimiter: ",Dumper($line_delimiter) if ($line_delimiter);
iterate_fld($cache,'display_data',$field,$page,$display_data,$x->{append});
}
$display_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;
+ }
}
}
- # fill data in index
- foreach my $d (@index_data) {
- $index->insert($field, $d, $path);
+ 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'}}) {
+ $x = unroll_x($x);
+
my $delimiter = x($x->{delimiter}) || ' ';
my $val = $cfg->val($database, x($x->{content}));
- 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");
- # no support for swish exact in config.
- # IMHO, it's useless
- ($s,$d,$i) = (0,0,1) if (lc($x->{type}) eq "index");
+ # 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);
+ $swish_data .= " ".$val if ($s);
$index->insert($field, $val, $path) if ($i);
}
#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 = $cache->{display_data}->{$field}->[$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);
$swish_data =~ s/ +/ /g;
$swish_data =~ s/ +$//g;
- $xml .= xmlify($field."_swish", unac_string($codepage,$swish_data));
+ $xml .= xmlify($field."_swish", my_unac_string($codepage,$swish_data));
}
my $swish_exact_data = $cache->{swish_exact_data}->{$field}->[$page];
# add delimiters before and after word.
# That is required to produce exact match
- $xml .= xmlify($field."_swish_exact", unac_string($codepage,'xxbxx '.$swish_exact_data.' xxexx'));
+ $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";
}
}
$swish_data =~ s/ +/ /g;
$swish_data =~ s/ +$//g;
- $xml .= xmlify($field."_swish", unac_string($codepage,$swish_data));
+ $xml .= xmlify($field."_swish", my_unac_string($codepage,$swish_data));
}
if ($swish_exact_data) {
# add delimiters before and after word.
# That is required to produce exact match
- $xml .= xmlify($field."_swish_exact", unac_string($codepage,'xxbxx '.$swish_exact_data.' xxexx'));
+ $xml .= xmlify($field."_swish_exact", my_unac_string($codepage,$swish_exact_data));
}
}
}
my $show_progress = $cfg_global->val('global', 'show_progress');
-my $unac_filter = $cfg_global->val('global', 'unac_filter');
-if ($unac_filter) {
- require $unac_filter;
+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 $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};
}
# open existing lookup file
my $type_base = $type;
$type_base =~ s/_.+$//g;
- $config=XMLin("./import_xml/$type.xml", forcearray => [ $type2tag{$type_base}, 'config', 'format' ], forcecontent => 1);
+ $config=XMLin("./import_xml/$type.xml", ForceArray => [ $type2tag{$type_base}, 'config', 'format' ], ForceContent => 1 );
+
+ # helper for progress bar
+ sub fmt_time {
+ my $t = shift || 0;
+ my $out = "";
+
+ my ($ss,$mm,$hh) = gmtime($t);
+ $out .= "${hh}h" if ($hh);
+ $out .= sprintf("%02d:%02d", $mm,$ss);
+ $out .= " " if ($hh == 0);
+ return $out;
+ }
# output current progress indicator
my $last_p = 0;
+ my $start_t = time();
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).">", $p );
+ if ($p < $last_p) {
+ $start_t = time();
+ } elsif ($p != $last_p) {
+ my $rate = ($current / (time() - $start_t || 1));
+ my $eta = ($total-$current) / ($rate || 1);
+ printf STDERR ("%5d [%-38s] %-5d %0.1f/s %s\r",$current,"=" x ($p/3)."$p%>", $total, $rate, fmt_time($eta));
$last_p = $p;
}
}
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 $isis_db\n";
- next;
- }
- }
-
- # OpenIsis::ERR_BADF
- if ($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 $db = new IsisDB( isisdb => $isis_db );
- 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 );
- }
+ my $max_rowid = $db->{'maxmfn'} || die "can't find maxmfn";
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}) {
-
+ my $row = $db->to_hash( $row_id );
+ if ($row) {
+
+ $row->{mfn} = $row_id;
+ $row->{record} = $db->{record};
+
progress($row->{mfn}, $max_rowid);
my $swishpath = $path."#".int($row->{mfn});
}
}
}
- # 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") {
- use Spreadsheet::ParseExcel;
- use Spreadsheet::ParseExcel::Utility qw(int2col);
+ 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!";
for(my $iC = $oWorksheet->{MinCol} ; defined $oWorksheet->{MaxCol} && $iC <= $oWorksheet->{MaxCol} ; $iC++) {
my $cell = $oWorksheet->{Cells}[$iR][$iC];
if ($cell) {
- $row->{int2col($iC)} = $cell->Value;
+ # 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) || $cell->Value;
}
}
}
} elsif ($type_base eq "marc") {
- use MARC;
+ require MARC::File::USMARC;
$import2cp = Text::Iconv->new($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->{format}) || 'usmarc';
-
+ warn "marc_format is no longer used!" if ($config->{marc_format});
print STDERR "Reading MARC file '$marc_file'\n";
- my $marc = new MARC;
- my $nr = $marc->openmarc({
- file=>$marc_file, format=>$format
- }) || die "Can't open MARC file '$marc_file'";
+ my $marc = MARC::File::USMARC->in( $marc_file )
+ || die "Can't open MARC file '$marc_file': ".$MARC::File::ERROR;
- my $i=0; # record nr.
+ # count records in MARC file
+ sub marc_count {
+ my $filename = shift || die;
+ my $file = MARC::File::USMARC->in($filename) || die $MARC::File::ERROR;
+ my $count = 0;
+ while ($file->skip()) {
+ $count++;
+ }
+ return $count;
+ }
+
+ my $count = marc_count($marc_file) || warn "no records in '$marc_file'?";
- my $rec;
+ my $i = 0;
- while ($marc->nextmarc(1)) {
+ while( my $rec = $marc->next() ) {
- # XXX
- fakeprogress($i++);
+ progress($i++,$count);
my $swishpath = $database."#".$i;
- if (my $xml = data2xml($type_base,$marc,$add_xml,$cfg,$database)) {
+ if (my $xml = data2xml($type_base,$rec,$add_xml,$cfg,$database)) {
$xml = $cp2utf->convert($xml);
use bytes; # as opposed to chars
print "Path-Name: $swishpath\n";
print "Document-Type: XML\n\n$xml\n";
}
}
+
+ print STDERR "\n";
+
} elsif ($type_base eq "feed") {
$import2cp = Text::Iconv->new($config->{feed_codepage},$codepage);
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, MARC
+This command will read ISIS data file using IsisDB perl module, MARC
records using MARC module and optionally Micro$oft Excel files to
create one XML file for usage with I<SWISH-E> indexer. Dispite it's name,
this script B<isn't general xml generator> 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<all2xml.conf>.
+
=head1 BUGS
Documentation is really lacking. However, in true Open Source spirit, source
projects / webpac / blobdiff