NAME

Normalize::Text::Music_Fields - normalize names of people's and (musical) works.

SYNOPSIS

   $name = $obj->Normalize::Text::Music_Fields::normalize_person($name);
   $work = $obj->Normalize::Text::Music_Fields::normalize_piece($work);
 # $obj should have methods `name_for_field_normalization', 'shorted_person'

DESCRIPTION

Databases of names and of works-per-name are taken from plain-text files (optionally in mail-header format). Names are stored in *.lst files. Works are stored in .comp files named after the shortened name of the composer.

The directories of these files are looked in the environment variable MUSIC_FIELDS_PATH (if defined, split the same way as PATH), or in $ENV{HOME}/.music_fields, and - (and - is replaced by the directory named as the module file with .pm dropped). At runtime, one can replace the list by calling function Normalize::Text::Music_Fields::set_path() with the list of directories as the argument.

(Since parsed files are cached, replacing the directory list should be done as early as possible.)

Files may be managed with utility subroutines provided with the module:

 # Translate from one-per-line to mail-header format:
 perl -wple "BEGIN {print q(# format = mail-header)} s/#\s*normalized\s*$//; $_ = qq(Title: $_) unless /^\s*(#|$)/; $_ = qq(\n$_) if $p and not /^##/; $_ .= qq(\n) unless $p = /^##/" Normalize::Text::Music_Fields-G_Gershwin.comp >Music_Fields-G_Gershwin.comp-mail

 # (inverse transformation:) Dump pieces listed in mail-header format
 perl -MNormalize::Text::Music_Fields -wle "print for Normalize::Text::Music_Fields::read_composer_file(shift, shift)" gershwin Music_Fields-G_Gershwin.comp-mail > o

 # Normalize data in 1-line-per piece format
 perl -MNormalize::Text::Music_Fields -wle "Normalize::Text::Music_Fields::prepare_tag_object_comp(shift)->Normalize::Text::Music_Fields::normalize_file_lines(shift)"

 # Create a mail-header file from a semi-processed (with "bold" fields)
 # mail-header file (with xml escapes, preceded by opus number)
 perl -MNormalize::Text::Music_Fields -00wnle "BEGIN {$tag = Normalize::Text::Music_Fields::prepare_tag_object_comp(shift @ARGV); print q(# format = mail-header)} print Normalize::Text::Music_Fields::emit_as_mail_header($tag,$_, q(bold,xml,opus),$pre)" shostakovich  o-xslt-better >Music_Fields-D_Shostakovich.comp-mail1

 # Likewise, from work-per-line with opus-numbers:
 perl -MNormalize::Text::Music_Fields -wnle "BEGIN {$tag = Normalize::Text::Music_Fields::prepare_tag_object_comp(shift @ARGV); print qq(# format = mail-header\n)} print Normalize::Text::Music_Fields::emit_as_mail_header($tag,$_, q(opus), $pre)" schnittke  o-schnittke-better  >Music_Fields-A_Schnittke.comp-mail2

 # A primitive tool for merging additional info into the database:
 perl -MNormalize::Text::Music_Fields -wnle "BEGIN {$tag = Normalize::Text::Music_Fields::prepare_tag_object_comp(shift @ARGV); print qq(# format = mail-header\n)} next unless s/^\s*\+\+\s*//; print Normalize::Text::Music_Fields::merge_info($tag,$_, q(opus,xml), qr(^(58|70|76|116|118|119)($|-)))" brahms o-brahms-op-no1-xslt

 # Minimal consistency check of persons database.
 perl -MNormalize::Text::Music_Fields -wle "BEGIN{binmode $_, ':encoding(cp866)' for \*STDIN, \*STDOUT, \*STDERR} print Normalize::Text::Music_Fields->check_persons"

 # Minimal testing code:
 perl -MNormalize::Text::Music_Fields -e Normalize::Text::Music_Fields::test_normalize_piece

It may be easier to type these examples if one uses manage_M_N_F.pm, which exports the mentioned subroutines to the main namespace (available in examples directory of a distribution of MP3::Tag). E.g., the last example becomes:

 perl -Mmanage_M_N_F -e test_normalize_piece