use Getopt::Long;
use Term::ReadLine;
-$| = 1;
-
my $term = new Term::ReadLine 'yaz-cleanup';
my $OUT = $term->OUT || \*STDOUT;
initialize($conf);
populate_trash() if ($conf->{trash});
-my @record = (); # current record storage
-my %recmeta = (); # metadata about current record
-my $recptr = 0; # record index pointer
-
-my $input = shift || 'incoming.marc.xml';
-
-open MARC, '<:utf8', $input;
+# 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";
print $NUMARC '<collection xmlns="http://www.loc.gov/MARC21/slim">',"\n";
-open my $EXMARC, '>:utf8', $conf->{exception};
+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,
}
do_automated_cleanups();
+ stow_record_data();
- $recptr = 0;
- until ($recptr == $#record) {
+ $ptr = 0;
+ until ($ptr == $#record) {
# naked ampersands
- if ($record[$recptr] =~ /&/ && $record[$recptr] !~ /&\w+?;/)
- { edit("Naked ampersand"); $recptr= 0; next }
+ if ($record[$ptr] =~ /&/ && $record[$ptr] !~ /&\w+?;/)
+ { edit("Naked ampersand"); $ptr= 0; next }
- # tags must be numeric
- if ($record[$recptr] =~ /<datafield tag="(.+?)"/) {
+ if ($record[$ptr] =~ /<datafield tag="(.+?)"/) {
my $match = $1;
+ # tags must be numeric
if ($match =~ /\D/) {
- edit("Non-numerics in tag");
- $recptr = 0;
+ 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[$recptr] =~ /<subfield code="(.*?)"/) {
+ if ($record[$ptr] =~ /<subfield code="(.*?)"/) {
my $match = $1;
if ($match =~ /\P{IsAlnum}/ or $match eq '') {
edit("Junk in subfield code/Null subfield code");
- $recptr = 0;
next;
}
}
- $recptr++;
+
+ $ptr++;
}
write_record($NUMARC);
}
#-----------------------------------------------------------------------------------
sub do_automated_cleanups {
- $recptr = 0;
- until ($recptr == $#record) {
+ $ptr = 0;
+ until ($ptr == $#record) {
# catch empty datafield elements
- if ($record[$recptr] =~ m/<datafield tag="..." ind1="." ind2=".">/) {
- if ($record[$recptr + 1] =~ m|</datafield>|) {
- my @a = @record[0 .. $recptr - 1];
- my @b = @record[$recptr + 2 .. $#record];
+ if ($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");
- $recptr = 0;
+ $ptr = 0;
next;
}
}
# and quasi-empty subfields
- if ($record[$recptr] =~ m|<subfield code="(.*?)">(.*?)</sub|) {
+ if ($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 .. $recptr - 1];
- my @b = @record[$recptr + 1 .. $#record];
+ my @a = @record[0 .. $ptr - 1];
+ my @b = @record[$ptr + 1 .. $#record];
@record = (@a, @b);
message("Empty subfield scrubbed");
- $recptr = 0;
+ $ptr = 0;
next;
}
}
- $recptr++;
+ $ptr++;
}
# single-line fixes
- for $recptr (0 .. $#record) {
+ for $ptr (0 .. $#record) {
# pad short leaders
- if ($record[$recptr] =~ m|<leader>(.+?)</leader>|) {
+ if ($record[$ptr] =~ m|<leader>(.+?)</leader>|) {
my $leader = $1;
if (length $leader < 24) {
$leader .= ' ' x (20 - length($leader));
$leader .= "4500";
- $record[$recptr] = "<leader>$leader</leader>\n";
+ $record[$ptr] = "<leader>$leader</leader>\n";
message("Short leader padded");
}
}
- if ($record[$recptr] =~ m|<controlfield tag="008">(.+?)</control|) {
+ if ($record[$ptr] =~ m|<controlfield tag="008">(.+?)</control|) {
#pad short 008
my $content = $1;
if (length $content < 40) {
$content .= ' ' x (40 - length($content));
- $record[$recptr] = "<controlfield tag=\"008\">$content</controlfield>\n";
+ $record[$ptr] = "<controlfield tag=\"008\">$content</controlfield>\n";
message("Short 008 padded");
}
}
# clean misplaced dollarsigns
- if ($record[$recptr] =~ m|<subfield code="\$">c?\d+\.\d{2}|) {
- $record[$recptr] =~ s|"\$">c?(\d+\.\d{2})|"c">\$$1|;
+ 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
- $record[$recptr] =~ s/tag=" /tag="00/g;
- $record[$recptr] =~ s/tag=" /tag="0/g;
- $record[$recptr] =~ s/tag="-/tag="0/g;
- $record[$recptr] =~ s/tag="(\d\d) /tag="0$1/g;
-
- # stow tag data if we're looking at it
- if ($record[$recptr] =~ m/<datafield tag="(.{3})" ind1="(.)" ind2="(.)">/) {
- $recmeta{tag} = $1;
- $recmeta{ind1} = $2;
- $recmeta{ind2} = $3;
- }
+ $record[$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
- $record[$recptr] =~ s/code=" ">c/code="c">/;
- $record[$recptr] =~ s/code=" ">\$/code="c"$>/;
+ $record[$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="$osub">(.+?)</);
+ }
+ }
}
}
# driver routines
#-----------------------------------------------------------------------------------
-
=head2 edit
Handles the Term::ReadLine loop
my ($msg) = @_;
return if $conf->{trash}{ $recmeta{tag} };
- message($msg, 1);
- print_context();
+ $conf->{editmsg} = $msg;
+ print_fullcontext();
+
+ # stow original problem line
+ $recmeta{origline} = $record[$ptr];
while (1) {
my $line = $term->readline('marc-cleanup>');
my @chunks = split /\s+/, $line;
- if (length $chunks[0] == 1)
- { next unless (defined $commands{$chunks[0]}) }
+ # 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 {
- $record[$recptr] = "$line\n";
- print_linecontext();
+ $recmeta{prevline} = $record[$ptr];
+ $record[$ptr] = "$line\n";
+ print_context();
}
}
+ # set pointer to top on the way out
+ $ptr = 0;
}
=head2 buildrecord
@record = @trimmed;
}
- # scrub newlines
- unless ($conf->{nocollapse}) {
- s/\n// for (@record);
- }
-
# add 903(?) with new record id
my $renumber = '';
if ($conf->{'renumber-from'}) {
- $renumber = join('', '<datafield tag="', $conf->{'renumber-tag'},
- '" ind1=" " ind2=" ">',
- '<subfield code="', $conf->{'renumber-subfield'}, '">',
- $conf->{'renumber-from'}, '</subfield></datafield>');
- $renumber .= "\n" if $conf->{nocollapse};
- push @record, $renumber;
+ $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 $FH "</record>\n";
}
-sub print_context {
- print " Tag:",$recmeta{tag}, " Ind1:'",
- $recmeta{ind1},"' Ind2:'", $recmeta{ind2}, "'\n";
- print_linecontext();
+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_linecontext {
- my $low = ($recptr - 3 < 0) ? 0 : $recptr - 3;
- print $OUT ' |', $record[$_] for ($low .. $recptr - 1);
- print $OUT '==> |', $record[$recptr];
- print $OUT ' |', $record[$recptr + 1], "\n";
+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, $new) = @_;
- print $OUT "\r", ' ' x 72, "\n" if $new;
+ my ($msg) = @_;
print $OUT "\r$msg at ",$conf->{ricount},"/",$conf->{rocount} + 1,"\n";
-
}
#-----------------------------------------------------------------------------------
sub substitute {
my ($line_in, @chunks) = @_;
+
my $ofrom = shift @chunks;
if ($ofrom =~ /^'/ or !@chunks) {
until ($ofrom =~ /'$/)
$char = "\\" . $char if ($char =~ /\W/);
$from = join('', $from, $char);
}
- $record[$recptr] =~ s/$from/$to/;
- print_linecontext();
+
+ $recmeta{prevline} = $record[$ptr];
+ $record[$ptr] =~ s/$from/$to/;
+ print_context();
return 0;
}
sub merge_lines {
+ $recmeta{prevline} = $record[$ptr];
# remove <subfield stuff; extract (probably wrong) subfield code
- $record[$recptr] =~ s/^\s*<subfield code="(.*?)">//;
+ $record[$ptr] =~ s/^\s*<subfield code="(.*?)">//;
# and move to front of line
- $record[$recptr] = join(' ', $1 , $record[$recptr]);
+ $record[$ptr] = join(' ', $1 , $record[$ptr]);
# tear off trailing subfield tag from preceeding line
- $record[$recptr - 1] =~ s|</subfield>\n||;
+ $record[$ptr - 1] =~ s|</subfield>\n||;
# join current line onto preceeding line
- $record[$recptr - 1] = join('', $record[$recptr - 1], $record[$recptr]);
+ $record[$ptr - 1] = join('', $record[$ptr - 1], $record[$ptr]);
# erase current line
- my @a = @record[0 .. $recptr - 1];
- my @b = @record[$recptr + 1 .. $#record];
+ my @a = @record[0 .. $ptr - 1];
+ my @b = @record[$ptr + 1 .. $#record];
@record = (@a, @b);
- # move recorde pointer to previous line
- $recptr--;
- print_linecontext();
+ # move record pointer to previous line
+ prev_line();
+ print_context();
+ return 0;
+}
+
+sub flip_line {
+ unless ($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 kill_line {
- my @a = @record[0 .. $recptr - 1];
- my @b = @record[$recptr + 1 .. $#record];
+ $recmeta{killline} = $record[$ptr];
+ my @a = @record[0 .. $ptr - 1];
+ my @b = @record[$ptr + 1 .. $#record];
@record = (@a, @b);
- print_linecontext();
+ print_context();
+ return 0;
+}
+
+sub yank_line {
+ unless ($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;
}
return 1;
}
-sub commit_edit { return 1 }
+sub next_line {
+ $ptr++ unless ($ptr == $#record);;
+ print_context();
+ return 0;
+}
-sub show_original {
- my ($line_in) = @_;
- print $OUT "\n$line_in\n";
+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 again ('C' for brief context)
- k Kill indicated line (remove from record)
- m Merge indicated line with previous line
- o Show original line
- s Substitute ARG1 for ARG2 in indicated 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;
'output|o=s',
'nocollapse|n',
'renumber-from|rf=i',
- 'original-tag|ot=i',
'renumber-tag|rt=i',
- 'renumber-subfield|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} = 'incoming.cleaned.marc.xml' unless defined $c->{output};
- $c->{exception} = 'incoming.exception.marc.xml' unless defined $c->{exception};
+ $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);
- #for my $key ('runtype', 'tag', 'subfield', 'output', 'exception')
- # { push @missing, $key unless $c->{$key} }
- #if (@missing) {
- # print "Required option: ", join(', ', @missing), " missing!\n";
- # show_help();
- #}
}
sub show_help {
print <<HELP;
-Usage is: $0 [OPTIONS] <filelist>
+Usage is: marc-cleanup [OPTIONS] <filelist>
Options
- --output -o Cleaned MARCXML output filename (default: incoming.cleaned.marc.xml)
- --exception -x Exception (dumped records) MARCXML filename (incoming.exception.marc.xml)
+ --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;
}