X-Git-Url: http://git.rot13.org/?a=blobdiff_plain;f=tsv2xp-xml.pl;fp=tsv2xp-xml.pl;h=446d32eb6b6d9194209210687a55e9c3562f9a7e;hb=7a4f4e55b47e297a98396094db133122554ce72b;hp=ccae4759182abcc75a3665991282557c469e25f1;hpb=e8410aea3a6465a2a94ce6d0d364b1e3fbcedd30;p=eprints3-migration.git diff --git a/tsv2xp-xml.pl b/tsv2xp-xml.pl index ccae475..446d32e 100755 --- a/tsv2xp-xml.pl +++ b/tsv2xp-xml.pl @@ -69,10 +69,21 @@ while(<$tsv>) { $file .= $1 if $full_path =~ m/(\.\w+)$/; } + my $subject; + if ( $file =~ m/^(\d+)(\w+)$/ ) { + $subject = + $2 =~ m/bib/i ? 'IZBIB' : + $2 =~ m/[info]{4}/i ? 'IZDHI' : + $2 =~ m/muz/i ? 'IZMUZ' : + $2 =~ m/arh/i ? 'IZARH' : + die "unknown subject: $2"; + } + my $c = $col; $header2col->{'eprintsid'} = $c; $v[$c++] = $eprintsid++; $header2col->{'file'} = $c; $v[$c++] = $file; $header2col->{'full_path'} = $c; $v[$c++] = $full_path; + $header2col->{'subject'} = $c; $v[$c++] = $subject; my $eprints = $xml; while ( $eprints =~ s//interpolate($1)/seg ) { @@ -80,6 +91,7 @@ while(<$tsv>) { } $eprints =~ s{.+}{}s if ! $full_path; + $eprints =~ s{.+}{}s if ! $subject; print $eprints; }