removing miker-filter series; superceded by filter_record_ids, which has far less...

[migration-tools.git] / fingerprinter

diff --git a/fingerprinter b/fingerprinter
index 246252d..1fa1f73 100755 (executable)
--- a/fingerprinter
+++ b/fingerprinter
@@ -16,9 +16,7 @@ $| = 1;
 initialize($conf);
 
 open OF, '>', $conf->{output};
-binmode(OF, ':utf8');
 open XF, '>', $conf->{exception};
-binmode(XF, ':utf8');
 
 for my $file (@ARGV) {
     print XF "Processing $file\n";
@@ -75,10 +73,22 @@ sub populate_marc {
     # date1, date2
     my $my_008 = $record->field('008');
     $my_008 = $my_008->as_string() if ($my_008);
-    unless (defined $my_008 and length $my_008 == 40)
-      { print XF ">> Bad 008 length in rec ",$marc{id},"\n"; return \%marc }
-    $marc{date1} = substr($my_008,7,4) if ($my_008);
-    $marc{date2} = substr($my_008,11,4) if ($my_008); # UNUSED
+    if (defined $my_008) {
+        unless (length $my_008 == 40) {
+            $my_008 = $my_008 . (' ' x (40 - length($my_008)));
+            print XF ">> Short 008 padded to ",length($my_008)," at rec $count\n";
+        }
+        $marc{date1} = substr($my_008,7,4) if ($my_008);
+        $marc{date2} = substr($my_008,11,4) if ($my_008); # UNUSED
+    }
+    unless ($marc{date1} and $marc{date1} =~ /\d{4}/) {
+        my $my_260 = $record->field('260');
+        my $date1 = $my_260->subfield('c') if $my_260;
+        if (defined $date1 and $date1 =~ /\d{4}/) {
+            $marc{date1} = $date1;
+            print XF ">> using 260c as date1 at rec $count\n";
+        }
+    }
 
     # item_form
     if ( $marc{record_type} =~ /[gkroef]/ ) { # MAP, VIS
@@ -103,6 +113,17 @@ sub populate_marc {
         }
     }
 
+    # oclc
+    $marc{oclc} = [];
+    push @{ $marc{oclc} }, $record->field('001')->as_string()
+      if ($record->field('001') and $record->field('003') and
+          $record->field('003')->as_string() =~ /OCo{0,1}LC/);
+    for ($record->field('035')) {
+        my $oclc = $_->subfield('a');
+        push @{ $marc{oclc} }, $oclc
+          if (defined $oclc and $oclc =~ /\(OCoLC\)/ and $oclc =~/([0-9]+)/);
+    }
+
     # "Accompanying material" (300e)
     $marc{accomp} = $record->field('300')->subfield('e')
       if $record->field('300');
@@ -187,13 +208,14 @@ sub marc_isvalid {
 
 sub dump_fingerprints {
     my ($marc) = @_;
+    my $good_fp = 0;
 
     if ($conf->{runtype} eq "primary") {
         print OF join("\t",$marc->{id}, $marc->{item_form},
                           $marc->{date1}, $marc->{record_type},
                           $marc->{bib_lvl}, $marc->{title}), "\n";
     } else {
-        if ((scalar @{ $marc->{isbns} } > 0) && $marc->{pages}) {
+        if ((scalar @{ $marc->{isbns} } > 0) and $marc->{pages}) {
             # case a : isbn and pages
             foreach my $isbn ( @{ $marc->{isbns}} ) {
                 print OF join("\t", $marc->{id}, "case a",
@@ -202,20 +224,31 @@ sub dump_fingerprints {
                                   $marc->{bib_lvl}, $marc->{title},
                                   $isbn, $marc->{pages}), "\n";
             }
+            $good_fp = 1;
         }
 
+=pod
+
+Looks like case b isn't a terribly good idea most of the time -- it
+will generate many spurious matches in series books in particular. So
+for now, simply turning it off.
+
         if ($marc->{edition}) { # case b : edition
             print OF join("\t", $marc->{id}, "case b",
                               $marc->{item_form}, $marc->{date1},
                               $marc->{record_type}, $marc->{bib_lvl},
                               $marc->{title}, $marc->{edition}), "\n";
+            $good_fp = 1;
         }
 
+=cut
+
         if ($marc->{issn}) { # case c : issn
             print OF join("\t", $marc->{id}, "case c",
                               $marc->{item_form}, $marc->{date1},
                               $marc->{record_type}, $marc->{bib_lvl},
                               $marc->{title}, $marc->{issn}), "\n";
+            $good_fp = 1;
         }
 
         if ($marc->{lccn}) { # case d : lccn
@@ -223,24 +256,45 @@ sub dump_fingerprints {
                               $marc->{item_form}, $marc->{date1},
                               $marc->{record_type}, $marc->{bib_lvl},
                               $marc->{title}, $marc->{lccn}) ,"\n";
+            $good_fp = 1;
         }
 
         if ($marc->{accomp}) { # case e : accomp
-            print OF join("\t", $marc->{id}, "case d",
+            print OF join("\t", $marc->{id}, "case e",
                               $marc->{item_form}, $marc->{date1},
                               $marc->{record_type}, $marc->{bib_lvl},
                               $marc->{title}, $marc->{accomp}) ,"\n";
+            $good_fp = 1;
+        }
+
+        # case o: oclc
+        if (scalar @{$marc->{oclc} }) {
+            for (@{$marc->{oclc} }) {
+                print OF join("\t", $marc->{id}, "case o",
+                              $marc->{item_form}, $marc->{date1},
+                              $marc->{record_type}, $marc->{bib_lvl},
+                              $marc->{title}, $_, "\n");
+            }
+            $good_fp = 1;
         }
 
         # case z : author, publisher, pubyear, pages
         if ($marc->{author} and $marc->{publisher} and $marc->{pubyear}
             and $marc->{pages}) {
-            print OF join("\t", $marc->{id}, "case e",
+            print OF join("\t", $marc->{id}, "case z",
                               $marc->{item_form}, $marc->{date1},
                               $marc->{record_type}, $marc->{bib_lvl},
                               $marc->{title}, $marc->{author},
                               $marc->{publisher}, $marc->{pubyear},
                               $marc->{pages}), "\n";
+            $good_fp = 1;
+        }
+
+        # case poo : nothing good; dump a primary and move on
+        unless ($good_fp) {
+            print OF join("\t", $marc->{id}, "case poo", $marc->{item_form},
+                          $marc->{date1}, $marc->{record_type},
+                          $marc->{bib_lvl}, $marc->{title}), "\n";
         }
     }
 }