use Getopt::Long;
use Term::ReadLine;
-$| = 1;
-
my $term = new Term::ReadLine 'yaz-cleanup';
my $OUT = $term->OUT || \*STDOUT;
-my $count = 0;
-my $reccount = 0;
-my $line = '';
-
-my @record = (); # current record storage
-my %reccontext = ();
-my @linecontext= (); # last 5 lines of file
-
-my $input = shift || 'incoming.marc.xml';
-
-open MARC, '<', $input;
-open my $NUMARC, '>', 'incoming.clean.marc.xml';
+# initialization and setup
+my $conf = {};
+initialize($conf);
+populate_trash() if ($conf->{trash});
+
+# set up files, since everything appears to be in order
+open MARC, '<:utf8', (shift || 'incoming.marc.xml');
+open my $NUMARC, '>:utf8', $conf->{output};
+open my $EXMARC, '>:utf8', $conf->{exception};
+open my $OLD2NEW, '>', 'old2new.incoming.map'
+ if ($conf->{'renumber-from'} and $conf->{'original-subfield'});
print $NUMARC '<?xml version="1.0" encoding="UTF-8"?>',"\n";
-open my $EXMARC, '>', 'incoming.exceptions.marc.xml';
-print $EXMARC '<?xml version="1.0" encoding="UTF-8"?>',"\n";
-open MARC2, '<', $input;
-<MARC2>;
+print $NUMARC '<collection xmlns="http://www.loc.gov/MARC21/slim">',"\n";
+
+my @record = (); # current record storage
+my %recmeta = (); # metadata about current record
+my $ptr = 0; # record index pointer
# this is the dispatch table which drives command selection in
# edit(), below
-my %commands = ( c => \&print_context,
- C => \&print_linecontext,
+my %commands = ( c => \&print_fullcontext,
+ n => \&next_line,
+ p => \&prev_line,
+ '<' => \&widen_window,
+ '>' => \&narrow_window,
+ d => \&display_lines,
+ o => \&insert_original,
k => \&kill_line,
- o => \&show_original,
+ y => \&yank_line,
+ f => \&flip_line,
m => \&merge_lines,
+ s => \&substitute,
t => \&commit_edit,
x => \&dump_record,
q => \&quit,
my @spinner = qw(- / | \\);
my $sidx = 0;
-while (my $line = getline()) {
- unless ($count % 2000) {
+while ( buildrecord() ) {
+ unless ($conf->{ricount} % 100) {
print "\rWorking... ", $spinner[$sidx];
$sidx = ($sidx == $#spinner) ? 0 : $sidx + 1;
}
- update_linecontext();
- # catch empty datafield elements
- if ($line =~ m|</datafield>|) {
- if ($record[-2] =~ m/<datafield tag="..." ind1="." ind2=".">/) {
- pop @record; pop @record;
- print $OUT "\rEmpty datafield scrubbed at line $count\n";
- next;
+ do_automated_cleanups();
+ stow_record_data();
+
+ $ptr = 0;
+ until ($ptr == $#record) {
+ # naked ampersands
+ if ($record[$ptr] =~ /&/ && $record[$ptr] !~ /&\w+?;/)
+ { edit("Naked ampersand"); $ptr= 0; next }
+
+ if ($record[$ptr] =~ /<datafield tag="(.+?)"/) {
+ my $match = $1;
+ # tags must be numeric
+ if ($match =~ /\D/) {
+ edit("Non-numerics in tag") unless $conf->{autoscrub};
+ next;
+ }
+ # test for existing 901/903 unless we're autocleaning them
+ unless ($conf->{'strip-nines'}) {
+ if ($match == 901 or $match == 903) {
+ edit("Incoming 901/903 found in data");
+ next;
+ }
+ }
+ }
+
+ # subfields can't be non-alphanumeric
+ if ($record[$ptr] =~ /<subfield code="(.*?)"/) {
+ my $match = $1;
+ if ($match =~ /\P{IsAlnum}/ or $match eq '') {
+ edit("Junk in subfield code/Null subfield code");
+ next;
+ }
}
- }
- # clean misplaced dollarsigns
- if ($line =~ m|<subfield code="\$">c?\d+\.\d{2}|) {
- $line =~ s|"\$">c?(\d+\.\d{2})|"c">\$$1|;
- print $OUT "\rDollar sign in subfield code corrected at line $count\n";
+ $ptr++;
}
+ write_record($NUMARC);
+}
+print $NUMARC "</collection>\n";
+print $OUT "\nDone. \n";
- # clean up tags with spaces in them
- $line =~ s/tag=" /tag="00/g;
- $line =~ s/tag=" /tag="0/g;
- $line =~ s/tag="-/tag="0/g;
- $line =~ s/tag="(\d\d) /tag="0$1/g;
- # stow tag data if we're looking at it
- if ($line =~ m/<datafield tag="(\d{3})" ind1="(.)" ind2="(.)">/) {
- $reccontext{tag} = $1;
- $reccontext{ind1} = $2;
- $reccontext{ind2} = $3;
+#-----------------------------------------------------------------------------------
+# cleanup routines
+#-----------------------------------------------------------------------------------
+
+sub do_automated_cleanups {
+ $ptr = 0;
+ until ($ptr == $#record) {
+ # catch empty datafield elements
+ if ($record[$ptr] =~ m/<datafield tag="..." ind1="." ind2=".">/) {
+ if ($record[$ptr + 1] =~ m|</datafield>|) {
+ my @a = @record[0 .. $ptr - 1];
+ my @b = @record[$ptr + 2 .. $#record];
+ @record = (@a, @b);
+ message("Empty datafield scrubbed");
+ $ptr = 0;
+ next;
+ }
+ }
+ # and quasi-empty subfields
+ if ($record[$ptr] =~ m|<subfield code="(.*?)">(.*?)</sub|) {
+ my $code = $1; my $content = $2;
+ if ($code =~ /\W/ and ($content =~ /\s+/ or $content eq '')) {
+ my @a = @record[0 .. $ptr - 1];
+ my @b = @record[$ptr + 1 .. $#record];
+ @record = (@a, @b);
+ message("Empty subfield scrubbed");
+ $ptr = 0;
+ next;
+ }
+ }
+ $ptr++;
}
- # naked ampersands
- if ($line =~ /&/ && $line !~ /&\w+?;/)
- { edit("Looks like naked ampersand", $line); next }
+ # single-line fixes
+ for $ptr (0 .. $#record) {
+ # pad short leaders
+ if ($record[$ptr] =~ m|<leader>(.+?)</leader>|) {
+ my $leader = $1;
+ if (length $leader < 24) {
+ $leader .= ' ' x (20 - length($leader));
+ $leader .= "4500";
+ $record[$ptr] = "<leader>$leader</leader>\n";
+ message("Short leader padded");
+ }
+ }
+ if ($record[$ptr] =~ m|<controlfield tag="008">(.+?)</control|) {
+ #pad short 008
+ my $content = $1;
+ if (length $content < 40) {
+ $content .= ' ' x (40 - length($content));
+ $record[$ptr] = "<controlfield tag=\"008\">$content</controlfield>\n";
+ message("Short 008 padded");
+ }
+ }
- # subfields can't be non-alphanumeric
- if ($line =~ /<subfield code="(.+?)"/) {
- my $match = $1;
- if ($match =~ /\P{IsAlnum}/) {
- edit("Junk in subfield", $line);
- next;
+ # clean misplaced dollarsigns
+ if ($record[$ptr] =~ m|<subfield code="\$">c?\d+\.\d{2}|) {
+ $record[$ptr] =~ s|"\$">c?(\d+\.\d{2})|"c">\$$1|;
+ message("Dollar sign corrected");
}
+
+ # clean up tags with spaces in them
+ $record[$ptr] =~ s/tag=" /tag="00/g;
+ $record[$ptr] =~ s/tag=" /tag="0/g;
+ $record[$ptr] =~ s/tag="-/tag="0/g;
+ $record[$ptr] =~ s/tag="(\d\d) /tag="0$1/g;
+
+ # automatable subfield maladies
+ $record[$ptr] =~ s/code=" ">c/code="c">/;
+ $record[$ptr] =~ s/code=" ">\$/code="c"$>/;
}
+}
+sub stow_record_data {
+ # get tag data if we're looking at it
+ if ($record[$ptr] =~ m/<datafield tag="(.{3})" ind1="(.)" ind2="(.)">/) {
+ $recmeta{tag} = $1;
+ $recmeta{ind1} = $2;
+ $recmeta{ind2} = $3;
+
+ # and since we are looking at a tag, see if it's the original id
+ if ($conf->{'original-subfield'} and $1 == $conf->{'original-tag'}) {
+ my $line = $record[$ptr]; my $lptr = $ptr;
+ my $osub = $conf->{'original-subfield'};
+ $recmeta{oid} = 'NONE';
+
+ until ($line =~ m|</record>|) {
+ $lptr++;
+ $line = $record[$lptr];
+ $recmeta{oid} = $1
+ if ($line =~ /<subfield code="$osub">(.+?)</);
+ }
+ }
+ }
}
-print $NUMARC "</xml>\n";
-print $EXMARC "</xml>\n";
+
+#-----------------------------------------------------------------------------------
+# driver routines
+#-----------------------------------------------------------------------------------
=head2 edit
=cut
sub edit {
- my ($msg, $line_in) = @_;
- print $OUT "\r".$msg, " at line $count:\n";
- print_context();
+ my ($msg) = @_;
+
+ return if $conf->{trash}{ $recmeta{tag} };
+ $conf->{editmsg} = $msg;
+ print_fullcontext();
+
+ # stow original problem line
+ $recmeta{origline} = $record[$ptr];
+
while (1) {
- my $line = $term->readline('yaz-cleanup>');
- if (length $line < 2)
- { next unless (defined $commands{$line}) }
- if (defined $commands{$line}) {
- my $term = $commands{$line}->($line_in);
+ my $line = $term->readline('marc-cleanup>');
+ my @chunks = split /\s+/, $line;
+
+ # lines with single-character first chunks are commands.
+ # make sure they exist.
+ if (length $chunks[0] == 1) {
+ unless (defined $commands{$chunks[0]}) {
+ print $OUT "No such command '", $chunks[0], "'\n";
+ next;
+ }
+ }
+
+ if (defined $commands{$chunks[0]}) {
+ my $term = $commands{$chunks[0]}->(@chunks[1..$#chunks]);
last if $term;
} else {
- if ($linecontext[3] eq " [LINE KILLED]\n") {
- push @record, "$line\n"
- } else {
- $record[-1] = "$line\n";
- }
- $linecontext[3] = "$line\n";
- print_linecontext();
+ $recmeta{prevline} = $record[$ptr];
+ $record[$ptr] = "$line\n";
+ print_context();
}
}
+ # set pointer to top on the way out
+ $ptr = 0;
}
-=head2 getline
+=head2 buildrecord
-Reads from the incoming MARC file; returns lines into the driver
-loop. Batches records for output, and maintains the linecontext listing.
+Constructs record arrays from the incoming MARC file and returns them
+to the driver loop.
=cut
-sub getline {
- my $l = <MARC>;
- $count++;
- if (defined $l) {
- if ($l =~ /<record>/) {
- @record = ($l);
- %reccontext = ();
- $reccount++;
- } elsif ($l =~ m|</record>|) {
- push @record, $l;
- write_record($NUMARC) if $reccount;
- } else {
- push @record, $l;
- }
- }
- return $l;
+sub buildrecord {
+ my $l = '';
+ $l = <MARC> while (defined $l and $l !~ /<record>/);
+ return $l unless defined $l;
+ @record = ($l);
+ %recmeta = ();
+ $conf->{ricount}++;
+
+ until ($l =~ m|</record>|)
+ { push @record, $l; $l = <MARC>; }
+ push @record, $l;
+ return 1;
}
sub write_record {
my ($FH) = @_;
- print $FH '<collection xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.loc.gov/MARC21/slim http://www.loc.gov/standards/marcxml/schema/MARC21slim.xsd" xmlns="http://www.loc.gov/MARC21/slim">',"\n";
+ my $trash = $conf->{trash};
+
+ $conf->{rocount}++ if ($FH eq $NUMARC);
+ print $FH '<!-- ', $recmeta{explanation}, " -->\n"
+ if(defined $recmeta{explanation});
+
+ # excise unwanted tags
+ if (keys %{$trash} or $conf->{autoscrub}) {
+ my @trimmed = ();
+ my $istrash = 0;
+ for my $line (@record) {
+ if ($istrash) {
+ $istrash = 0 if $line =~ m|</datafield|;
+ next;
+ }
+ if ($line =~ m/<datafield tag="(.{3})"/) {
+ my $tag = $1;
+ if ($trash->{$tag} or ($conf->{autoscrub} and $tag =~ /\D/)) {
+ $istrash = 1;
+ next
+ }
+ }
+ push @trimmed, $line;
+ }
+ @record = @trimmed;
+ }
+
+ # add 903(?) with new record id
+ my $renumber = '';
+ if ($conf->{'renumber-from'}) {
+ $recmeta{nid} = $conf->{'renumber=from'};
+ $renumber = join('', ' <datafield tag="', $conf->{'renumber-tag'},
+ '" ind1=" " ind2=" "> <subfield code="',
+ $conf->{'renumber-subfield'},
+ '">', $recmeta{nid}, "</subfield></datafield>\n");
+ my @tmp = @record[0 .. $#record - 1];
+ my $last = $record[$#record];
+ @record = (@tmp, $renumber, $last);
+ $conf->{'renumber-from'}++;
+ }
+
+ # scrub newlines
+ unless ($conf->{nocollapse})
+ { s/\n// for (@record) }
+
+ # write to old->new map file if needed
+ print $OLD2NEW $recmeta{oid}, "\t", $recmeta{nid}, "\n"
+ if ($conf->{'renumber-from'} and $conf->{'original-subfield'});
+
+ # and finally, actually write the record
print $FH @record;
- print $FH "</collection>\n";
}
-sub update_linecontext {
- my $line2 = <MARC2>;
- push @linecontext, $line2;
- shift @linecontext if (@linecontext > 5);
+sub print_fullcontext {
+ print $OUT "\r", ' ' x 72, "\n";
+ print $OUT $conf->{editmsg},"\n";
+ print $OUT "\r Tag:",$recmeta{tag}, " Ind1:'",
+ $recmeta{ind1},"' Ind2:'", $recmeta{ind2}, "'";
+ print_context();
+ return 0;
+}
+
+sub print_context {
+ my $upper = int($conf->{window} / 2) + 1;
+ my $lower = int($conf->{window} / 2) - 1;
+ my $start = ($ptr - $upper < 0) ? 0 : $ptr - $upper;
+ my $stop = ($ptr + $lower > $#record) ? $#record : $ptr + $lower;
+ print $OUT "\n";
+ print $OUT ' |', $record[$_] for ($start .. $ptr - 1);
+ print $OUT '==> |', $record[$ptr];
+ print $OUT ' |', $record[$_] for ($ptr + 1 .. $stop);
+ print $OUT "\n";
+ return 0;
+}
+
+sub message {
+ my ($msg) = @_;
+ print $OUT "\r$msg at ",$conf->{ricount},"/",$conf->{rocount} + 1,"\n";
}
#-----------------------------------------------------------------------------------
# command routines
#-----------------------------------------------------------------------------------
-sub print_context {
- print "\n Tag:",$reccontext{tag}, " Ind1:'",
- $reccontext{ind1},"' Ind2:'", $reccontext{ind2}, "'";
- print_linecontext();
+sub substitute {
+ my ($line_in, @chunks) = @_;
+
+ my $ofrom = shift @chunks;
+ if ($ofrom =~ /^'/ or !@chunks) {
+ until ($ofrom =~ /'$/)
+ { $ofrom .= join(' ','',shift @chunks) }
+ $ofrom =~ s/^'//; $ofrom =~ s/'$//;
+ }
+ my $to = shift @chunks;
+ if ($to =~ /^'/) {
+ until ($to =~ /'$/ or !@chunks)
+ { $to .= join(' ','',shift @chunks) }
+ $to =~ s/^'//; $to =~ s/'$//;
+ }
+
+ my $from = '';
+ for my $char (split(//,$ofrom)) {
+ $char = "\\" . $char if ($char =~ /\W/);
+ $from = join('', $from, $char);
+ }
+
+ $recmeta{prevline} = $record[$ptr];
+ $record[$ptr] =~ s/$from/$to/;
+ print_context();
return 0;
}
-sub print_linecontext {
- print $OUT "\n", join(' |','',@linecontext[0..2]);
- print $OUT '==> |', $linecontext[3];
- print $OUT ' |', $linecontext[4],"\n";
+sub merge_lines {
+ $recmeta{prevline} = $record[$ptr];
+ # remove <subfield stuff; extract (probably wrong) subfield code
+ $record[$ptr] =~ s/^\s*<subfield code="(.*?)">//;
+ # and move to front of line
+ $record[$ptr] = join(' ', $1 , $record[$ptr]);
+ # tear off trailing subfield tag from preceeding line
+ $record[$ptr - 1] =~ s|</subfield>\n||;
+ # join current line onto preceeding line
+ $record[$ptr - 1] = join('', $record[$ptr - 1], $record[$ptr]);
+ # erase current line
+ my @a = @record[0 .. $ptr - 1];
+ my @b = @record[$ptr + 1 .. $#record];
+ @record = (@a, @b);
+ # move record pointer to previous line
+ prev_line();
+ print_context();
return 0;
}
-sub show_original {
- my ($line_in) = @_;
- print $OUT "\n$line_in\n";
+sub flip_line {
+ unless ($recmeta{prevline})
+ { print $OUT "No previously edited line to flip\n"; return }
+ my $temp = $record[$ptr];
+ $record[$ptr] = $recmeta{prevline};
+ $recmeta{prevline} = $temp;
+ print_context();
return 0;
}
-sub merge_lines {
- my $last = pop @record;
- $last =~ s/^\s+//;
- $record[-1] =~ s/\n//;
- $record[-1] = join('', $record[-1], $last);
- my @temp = ("\n");
- push @temp, @linecontext[0..1];
- $temp[3] = $record[-1];
- $temp[4] = $linecontext[4];
- @linecontext = @temp;
- print_linecontext();
+sub kill_line {
+ $recmeta{killline} = $record[$ptr];
+ my @a = @record[0 .. $ptr - 1];
+ my @b = @record[$ptr + 1 .. $#record];
+ @record = (@a, @b);
+ print_context();
return 0;
}
-sub kill_line {
- pop @record;
- $linecontext[3] = " [LINE KILLED]\n";
- print_linecontext();
+sub yank_line {
+ unless ($recmeta{killline})
+ { print $OUT "No killed line to yank\n"; return }
+ my @a = @record[0 .. $ptr - 1];
+ my @b = @record[$ptr .. $#record];
+ @record = (@a, $conf->{killline}, @b);
+ print_context();
+ return 0;
+}
+
+sub display_lines {
+ print $OUT "\nOrig. edit line :", $recmeta{origline};
+ print $OUT "Current flip line:", $recmeta{prevline} if $recmeta{prevline};
+ print $OUT "Last killed line :", $recmeta{killline} if $recmeta{killline};
+ print $OUT "\n";
return 0;
}
sub dump_record {
- my $line = <MARC>; $count++;
- update_linecontext();
- until ($line =~ m|</record>|) {
- push @record, $line;
- $line = <MARC>; $count++;
- update_linecontext();
- }
- push @record, $line;
+ my ($line_in, @explanation) = @_;
+ $recmeta{explanation} = join(' ', 'Tag', $recmeta{tag}, @explanation);
write_record($EXMARC);
return 1;
}
+sub next_line {
+ $ptr++ unless ($ptr == $#record);;
+ print_context();
+ return 0;
+}
+
+sub prev_line {
+ $ptr-- unless ($ptr == 0);
+ print_context();
+ return 0;
+}
+
sub commit_edit { return 1 }
+sub widen_window {
+ if ($conf->{window} == 15)
+ { print $OUT "Window can't be bigger than 15 lines\n"; return }
+ $conf->{window} += 2;
+ print_context;
+}
+
+sub narrow_window {
+ if ($conf->{window} == 5)
+ { print $OUT "Window can't be smaller than 5 lines\n"; return }
+ $conf->{window} -= 2;
+ print_context;
+}
+
sub help {
print $OUT <<HELP;
-
Type a replacement for the indicated line, or enter a command.
-Commands: c Show record context ('C' for brief context)
- k Kill this line (remove from record)
- m Merge indicated line with previous line
- o Show original line
- t Commit changes and resume stream edit
- x Write this record to the exception file instead of output
- q Quit
+DISPLAY COMMANDS | LINE AUTO-EDIT COMMANDS
+< Expand context window | k Kill current line
+> Contract context window | y Yank last killed line
+p Move pointer to prev line | m Merge current line into preceding line
+n Move pointer to next line | o Insert original line
+c Print line context | f Flip current line and last edited line
+d Print current saved lines |
+-----------------------------+-------------------------------------------
+s Subtitute; replace ARG1 in current line with ARG2. If either ARG
+ contains spaces, it must be single-quoted
+t Commit changes and resume automated operations
+x Dump record to exception file
+q Quit
HELP
return 0;
}
sub quit { exit }
+
+#-----------------------------------------------------------------------------------
+# populate_trash
+#-----------------------------------------------------------------------------------
+# defined a domain-specific language for specifying MARC tags to be dropped from
+# records during processing. it is line oriented, and is specified as follows:
+#
+# each line may specify any number of tags to be included, either singly (\d{1,3})
+# or as a range (\d{1,3}\.\.\d{1,3}
+#
+# if a single number is given, it must be between '000' and '999', inclusive.
+#
+# ranges obey the previous rule, and also the first number of the range must be less
+# than the second number
+#
+# finally, any single range in a line may be followed by the keyword 'except'. every
+# number or range after 'except' is excluded from the range specified. all these
+# numbers must actually be within the range.
+#
+# specifying a tag twice is an error, to help prevent typos
+
+sub populate_trash {
+ print $OUT ">>> TRASHTAGS FILE FOUND. LOADING TAGS TO BE STRIPPED FROM OUTPUT\n";
+ open TRASH, '<', $conf->{trash}
+ or die "Can't open trash tags file!\n";
+ while (<TRASH>) {
+ my $lastwasrange = 0;
+ my %lastrange = ( high => 0, low => 0);
+ my $except = 0;
+
+ my @chunks = split /\s+/;
+ while (my $chunk = shift @chunks) {
+
+ # single values
+ if ($chunk =~ /^\d{1,3}$/) {
+ trash_add($chunk, $except);
+ $lastwasrange = 0;
+ next;
+ }
+
+ # ranges
+ if ($chunk =~ /^\d{1,3}\.\.\d{1,3}$/) {
+ my ($low, $high) = trash_add_range($chunk, $except, \%lastrange);
+ $lastwasrange = 1;
+ %lastrange = (low => $low, high => $high)
+ unless $except;
+ next;
+ }
+
+ # 'except'
+ if ($chunk eq 'except') {
+ die "Keyword 'except' can only follow a range (line $.)\n"
+ unless $lastwasrange;
+ die "Keyword 'except' may only occur once per line (line $.)\n"
+ if $except;
+ $except = 1;
+ next;
+ }
+
+ die "Unknown chunk $chunk in .trashtags file (line $.)\n";
+ }
+ }
+
+ # remove original id sequence tag from trash hash if we know it
+ trash_add($conf->{'original-tag'}, 1)
+ if ($conf->{'original-tag'} and $conf->{trash}{ $conf->{'original-tag'} });
+}
+
+sub trash_add_range {
+ my ($chunk, $except, $range) = @_;
+ my ($low,$high) = split /\.\./, $chunk;
+ die "Ranges must be 'low..high' ($low is greater than $high on line $.)\n"
+ if ($low > $high);
+ if ($except) {
+ die "Exception ranges must be within last addition range (line $.)\n"
+ if ($low < $range->{low} or $high > $range->{high});
+ }
+ for my $tag ($low..$high) {
+ trash_add($tag, $except)
+ }
+ return $low, $high;
+}
+
+sub trash_add {
+ my ($tag, $except) = @_;
+ my $trash = $conf->{trash};
+
+ die "Trash values must be valid tags (000-999)\n"
+ unless ($tag >= 0 and $tag <= 999);
+
+ if ($except) {
+ delete $trash->{$tag};
+ } else {
+ die "Trash tag '$tag' specified twice (line $.)\n"
+ if $trash->{$tag};
+ $trash->{$tag} = 1;
+ }
+}
+
+#-----------------------------------------------------------------------
+
+=head2 initialize
+
+Performs boring script initialization. Handles argument parsing,
+mostly.
+
+=cut
+
+sub initialize {
+ my ($c) = @_;
+ my @missing = ();
+
+ # set mode on existing filehandles
+ binmode(STDIN, ':utf8');
+
+ my $rc = GetOptions( $c,
+ 'autoscrub|a',
+ 'exception|x=s',
+ 'output|o=s',
+ 'nocollapse|n',
+ 'renumber-from|rf=i',
+ 'renumber-tag|rt=i',
+ 'renumber-subfield|rs=i',
+ 'original-tag|ot=i',
+ 'original-subfield|os=i',
+ 'script',
+ 'strip-nines',
+ 'trash|t=s',
+ 'trashhelp',
+ 'help|h',
+ );
+ show_help() unless $rc;
+ show_help() if ($c->{help});
+ show_trashhelp() if ($c->{trashhelp});
+
+ # defaults
+ $c->{output} = 'cleaned.incoming.marc.xml' unless defined $c->{output};
+ $c->{exception} = 'cleaned.exceptions.marc.xml' unless defined $c->{exception};
+ $c->{'renumber-tag'} = 903 unless defined $c->{'renumber-tag'};
+ $c->{'renumber-subfield'} = 'a' unless defined $c->{'renumber-subfield'};
+ $c->{window} = 5;
+
+ # autotrash 901, 903 if strip-nines
+ if ($c->{'strip-nines'}) {
+ $c->{trash}{901} = 1;
+ $c->{trash}{903} = 1;
+ }
+
+ my @keys = keys %{$c};
+ show_help() unless (@ARGV and @keys);
+}
+
+sub show_help {
+ print <<HELP;
+Usage is: marc-cleanup [OPTIONS] <filelist>
+Options
+ --output -o Cleaned MARCXML output filename
+ (default: cleaned.incoming.marc.xml)
+ --exception -x Exception (dumped records) MARCXML filename
+ (cleaned.exceptions.marc.xml)
+ --trash File containing trash tag data (see --trashhelp)
+
+ --renumber-from=NUM -rf Begin renumbering id sequence with this number
+ --renumber-tag -rt Tag to use in renumbering (default: 903)
+ --renumber-subfield -rs Subfield code to use in renumbering (default: a)
+ --original-tag -ot Original id tag; will be kept in output even if
+ it appears in the trash file
+ --original-subfield -os Original id subfield code. If this is specified
+ and renumbering is in effect, an old-to-new mapping
+ file (old2new.incoming.map) will be generated.
+
+ --nocollapse -n Don't compress records to one line on output
+ --autoscrub -a Automatically remove non-numeric tags in data
+ --strip-nines Automatically remove any existing 901/903 tags in data
+
+ --script Store human-initiated ops in scriptfile (.mcscript)
+ Not yet implemented
+HELP
+exit;
+}
+
+sub show_trashhelp {
+ print <<HELP;
+The marc-cleanup trash tags file is a simple plaintext file. It is a
+line oriented format. There are three basic tokens:
+
+ * The tag
+ * The tag range
+ * The "except" clause
+
+Any number of tags and/or tag ranges can appear on a single line. A
+tag cannot appear twice in the file, either alone or as part of a
+range. This is to prevent errors in the trash tag listing. Items do
+not have to be sorted within a line. These following lines are valid:
+
+ 850 852 870..879 886 890 896..899
+ 214 696..699 012
+
+Ranges must be ordered internally. That is, "870..879" is valid while
+"879..870" is not.
+
+Finally, there can be only one "except" clause on a line. It is
+composed of the word "except" followed by one or more tags or
+ranges. Except clauses must follow a range, and all tags within the
+clause must be within the range which the clause follows.
+
+ 900..997 except 935 950..959 987 994
+
+is a valid example.
+HELP
+exit;
+}