sync w/ main trunk

[bioperl-live.git] / Bio / DB / Flat.pm

#
# $Id$
#
# BioPerl module for Bio::DB::Flat
#
# Please direct questions and support issues to <bioperl-l@bioperl.org> 
#
# Cared for by Lincoln Stein <lstein@cshl.org>
#
# You may distribute this module under the same terms as perl itself

# POD documentation - main docs before the code

=head1 NAME

Bio::DB::Flat - Interface for indexed flat files

=head1 SYNOPSIS

  $db = Bio::DB::Flat->new(-directory  => '/usr/share/embl',
                           -dbname     => 'mydb',
                           -format     => 'embl',
                           -index      => 'bdb',
                           -write_flag => 1);
  $db->build_index('/usr/share/embl/primate.embl',
                   '/usr/share/embl/protists.embl');
  $seq       = $db->get_Seq_by_id('BUM');
  @sequences = $db->get_Seq_by_acc('DIV' => 'primate');
  $raw       = $db->fetch_raw('BUM');

=head1 DESCRIPTION

This object provides the basic mechanism to associate positions in
files with primary and secondary name spaces. Unlike
Bio::Index::Abstract (see L<Bio::Index::Abstract>), this is specialized
to work with the "flat index" and BerkeleyDB indexed flat file formats
worked out at the 2002 BioHackathon.

This object is a general front end to the underlying databases.

=head1 FEEDBACK

=head2 Mailing Lists

User feedback is an integral part of the evolution of this and other
Bioperl modules. Send your comments and suggestions preferably to one
of the Bioperl mailing lists.  Your participation is much appreciated.

  bioperl-l@bioperl.org                  - General discussion
  http://bioperl.org/wiki/Mailing_lists  - About the mailing lists

=head2 Support 
 
Please direct usage questions or support issues to the mailing list:
  
L<bioperl-l@bioperl.org>
  
rather than to the module maintainer directly. Many experienced and 
reponsive experts will be able look at the problem and quickly 
address it. Please include a thorough description of the problem 
with code and data examples if at all possible.

=head2 Reporting Bugs

Report bugs to the Bioperl bug tracking system to help us keep track
the bugs and their resolution.  Bug reports can be submitted via the
web:

  http://bugzilla.open-bio.org/

=head1 AUTHOR - Lincoln Stein

Email - lstein@cshl.org

=head1 APPENDIX

The rest of the documentation details each of the object methods. Internal
methods are usually preceded with an "_" (underscore).

=cut


# Let the code begin...
package Bio::DB::Flat;

use File::Spec;

use base qw(Bio::Root::Root Bio::DB::RandomAccessI);

use constant CONFIG_FILE_NAME => 'config.dat';

=head2 new

 Title   : new
 Usage   : my $db = Bio::DB::Flat->new(
                     -directory  => $root_directory,
                     -dbname     => 'mydb',
                     -write_flag => 1,
                     -index      => 'bdb',
                     -verbose    => 0,
                     -out        => 'outputfile',
                     -format     => 'genbank');
 Function: create a new Bio::DB::Flat object
 Returns : new Bio::DB::Flat object
 Args    : -directory    Root directory containing "config.dat"
           -write_flag   If true, allows creation/updating.
           -verbose      Verbose messages
           -out          File to write to when write_seq invoked
           -index        'bdb' or 'binarysearch'
 Status  : Public

The required -directory argument indicates where the flat file indexes
will be stored.  The build_index() and write_seq() methods will
automatically create subdirectories of this root directory.  Each
subdirectory will contain a human-readable configuration file named
"config.dat" that specifies where the individual indexes are stored.

The required -dbname argument gives a name to the database index.  The
index files will actually be stored in a like-named subdirectory
underneath the root directory.

The -write_flag enables writing new entries into the database as well
as the creation of the indexes.  By default the indexes will be opened
read only.

-index is one of "bdb" or "binarysearch" and indicates the type of
index to generate.  "bdb" corresponds to Berkeley DB.  You *must* be
using BerkeleyDB version 2 or higher, and have the Perl BerkeleyDB
extension installed (DB_File will *not* work). "binarysearch"
corresponds to the OBDA "flat" indexed file.

The -out argument specifies the output file for writing objects created
with write_seq().

The -format argument specifies the format of the input file or files. If
the file suffix is one that Bioperl can already associate with a format
then this is optional.

=cut

sub new {
  my $class = shift;
  $class  = ref($class) if ref($class);
  my $self = $class->SUPER::new(@_);

  # first we initialize ourselves
  my ($flat_directory,$dbname,$format) = 
    $self->_rearrange([qw(DIRECTORY DBNAME FORMAT)],@_);

  defined $flat_directory
    or $self->throw('Please supply a -directory argument');
  defined $dbname
    or $self->throw('Please supply a -dbname argument');

  # set values from configuration file
  $self->directory($flat_directory);
  $self->dbname($dbname);

  $self->throw("Base directory $flat_directory doesn't exist")
    unless -e $flat_directory;
  $self->throw("$flat_directory isn't a directory")
    unless -d _;
  my $dbpath = File::Spec->catfile($flat_directory,$dbname);
  unless (-d $dbpath) {
    $self->debug("creating db directory $dbpath\n");
    mkdir $dbpath,0777 or $self->throw("Can't create $dbpath: $!");
  }
  $self->_read_config();

  # but override with initialization values
  $self->_initialize(@_);

  $self->throw('you must specify an indexing scheme') 
    unless $self->indexing_scheme;

  # now we figure out what subclass to instantiate
  my $index_type = $self->indexing_scheme eq 'BerkeleyDB/1' ? 'BDB'
                  :$self->indexing_scheme eq 'flat/1'       ? 'Binary'
                  :$self->throw("unknown indexing scheme: " .
                                $self->indexing_scheme);
  $format = $self->file_format;

  # because Michele and Lincoln did it differently
  # Michele's way is via a standalone concrete class
  if ($index_type eq 'Binary') {
    my $child_class = 'Bio::DB::Flat::BinarySearch';
    eval "use $child_class";
    $self->throw($@) if $@;
    push @_, ('-format', $format);
    return $child_class->new(@_);
  }

  # Lincoln uses Bio::SeqIO style delegation.
  my $child_class= "Bio\:\:DB\:\:Flat\:\:$index_type\:\:\L$format";
  eval "use $child_class";
  $self->throw($@) if $@;

  # rebless & reinitialize with the new class
  # (this prevents subclasses from forgetting to call our own initialization)
  bless $self,$child_class;
  $self->_initialize(@_);
  $self->_set_namespaces(@_);

  $self;
}

sub _initialize {
  my $self = shift;

  my ($flat_write_flag,$dbname,$flat_indexing,$flat_verbose,$flat_outfile,$flat_format)
    = $self->_rearrange([qw(WRITE_FLAG DBNAME INDEX VERBOSE OUT FORMAT)],@_);

  $self->write_flag($flat_write_flag) if defined $flat_write_flag;

  if (defined $flat_indexing) {
    # very permissive
    $flat_indexing = 'BerkeleyDB/1' if $flat_indexing =~ /bdb/;
    $flat_indexing = 'flat/1'       if $flat_indexing =~ /^(flat|binary)/;
    $self->indexing_scheme($flat_indexing);
  }

  $self->verbose($flat_verbose)    if defined $flat_verbose;
  $self->dbname($dbname)           if defined $dbname;
  $self->out_file($flat_outfile)   if defined $flat_outfile;
  $self->file_format($flat_format) if defined $flat_format;
}

sub _set_namespaces {
  my $self = shift;

  $self->primary_namespace($self->default_primary_namespace)
    unless defined $self->{flat_primary_namespace};

  $self->secondary_namespaces($self->default_secondary_namespaces)
    unless defined $self->{flat_secondary_namespaces};

  $self->file_format($self->default_file_format)
    unless defined $self->{flat_format};
}

=head2 new_from_registry

 Title   : new_from_registry
 Usage   : $db = Bio::DB::Flat->new_from_registry(%config)
 Function: creates a new Bio::DB::Flat object in a Bio::DB::Registry-
           compatible fashion
 Returns : new Bio::DB::Flat
 Args    : provided by the registry, see below
 Status  : Public

The following registry-configuration tags are recognized:

  location     Root of the indexed flat file; corresponds to the new() method's
               -directory argument.

=cut

sub new_from_registry {
   my ($self,%config) =  @_;
   my $location = $config{'location'} or 
     $self->throw('location tag must be specified.');
   my $dbname   = $config{'dbname'}   or 
     $self->throw('dbname tag must be specified.');

   my $db = $self->new(-directory => $location,
                        -dbname    => $dbname,
                      );
   $db;
}

# accessors
sub directory {
  my $self = shift;
  my $d = $self->{flat_directory};
  $self->{flat_directory} = shift if @_;
  $d;
}
sub write_flag {
  my $self = shift;
  my $d = $self->{flat_write_flag};
  $self->{flat_write_flag} = shift if @_;
  $d;
}
sub verbose {
  my $self = shift;
  my $d = $self->{flat_verbose};
  $self->{flat_verbose} = shift if @_;
  $d;
}
sub out_file {
  my $self = shift;
  my $d = $self->{flat_outfile};
  $self->{flat_outfile} = shift if @_;
  $d;
}
sub dbname {
  my $self = shift;
  my $d = $self->{flat_dbname};
  $self->{flat_dbname} = shift if @_;
  $d;
}
sub primary_namespace {
  my $self = shift;
  my $d    = $self->{flat_primary_namespace};
  $self->{flat_primary_namespace} = shift if @_;
  $d;
}

# get/set secondary namespace(s)
# pass an array ref.
# get an array ref in scalar context, list in list context.
sub secondary_namespaces {
  my $self = shift;
  my $d    = $self->{flat_secondary_namespaces};
  $self->{flat_secondary_namespaces} = (ref($_[0]) eq 'ARRAY' ? shift : [@_]) if @_;
  return unless $d;
  $d = [$d] if $d && ref($d) ne 'ARRAY';  # just paranoia
  return wantarray ? @$d : $d;
}

# return the file format
sub file_format {
  my $self = shift;
  my $d    = $self->{flat_format};
  $self->{flat_format} = shift if @_;
  $d;
}

# return the alphabet
sub alphabet {
  my $self = shift;
  my $d    = $self->{flat_alphabet};
  $self->{flat_alphabet} = shift if @_;
  $d;
}

sub parse_one_record {
  my $self  = shift;
  my $fh    = shift;
  my $parser =
    $self->{cached_parsers}{fileno($fh)}
      ||= Bio::SeqIO->new(-fh=>$fh,-format=>$self->default_file_format);
  my $seq = $parser->next_seq or return;
  $self->{flat_alphabet} ||= $seq->alphabet;
  my $ids = $self->seq_to_ids($seq);
  return $ids;
}


# return the indexing scheme
sub indexing_scheme {
  my $self = shift;
  my $d    = $self->{flat_indexing};
  $self->{flat_indexing} = shift if @_;
  $d;
}

sub add_flat_file {
  my $self = shift;
  my ($file_path,$file_length,$nf) = @_;

  # check that file_path is absolute
  unless (File::Spec->file_name_is_absolute($file_path)) {
    $file_path = File::Spec->rel2abs($file_path);
  }

  -r $file_path or $self->throw("flat file $file_path cannot be read: $!");

  my $current_size = -s _;
  if (defined $file_length) {
    $current_size == $file_length
      or $self->throw("flat file $file_path has changed size.  Was $file_length bytes; now $current_size");
  } else {
    $file_length = $current_size;
  }

  unless (defined $nf) {
    $self->{flat_file_index} = 0 unless exists $self->{flat_file_index};
    $nf = $self->{flat_file_index}++;
  }
  $self->{flat_flat_file_path}{$nf}      = $file_path;
  $self->{flat_flat_file_no}{$file_path} = $nf;
  $nf;
}

sub write_config {
  my $self = shift;
  $self->write_flag or $self->throw("cannot write configuration file because write_flag is not set");
  my $path = $self->_config_path;

  open (my $F,">$path") or $self->throw("open error on $path: $!");

  my $index_type = $self->indexing_scheme;
  print $F "index\t$index_type\n";

  my $format     = $self->file_format;
  my $alphabet   = $self->alphabet;
  my $alpha      = $alphabet ? "/$alphabet" : '';
  print $F "format\tURN:LSID:open-bio.org:${format}${alpha}\n";

  my @filenos = $self->_filenos or $self->throw("cannot write config file because no flat files defined");
  for my $nf (@filenos) {
    my $path = $self->{flat_flat_file_path}{$nf};
    my $size = -s $path;
    print $F join("\t","fileid_$nf",$path,$size),"\n";
  }

  # write primary namespace
  my $primary_ns = $self->primary_namespace
    or $self->throw('cannot write config file because no primary namespace defined');

  print $F join("\t",'primary_namespace',$primary_ns),"\n";

  # write secondary namespaces
  my @secondary = $self->secondary_namespaces;
  print $F join("\t",'secondary_namespaces',@secondary),"\n";

  close $F or $self->throw("close error on $path: $!");
}

sub files {
  my $self = shift;
  return unless $self->{flat_flat_file_no};
  return keys %{$self->{flat_flat_file_no}};
}

sub write_seq {
  my $self  = shift;
  my $seq   = shift;

  $self->write_flag or $self->throw("cannot write sequences because write_flag is not set");

  my $file  = $self->out_file or $self->throw('no outfile defined; use the -out argument to new()');
  my $seqio = $self->{flat_cached_parsers}{$file}
    ||= Bio::SeqIO->new(-Format => $self->file_format,
                        -file   => ">$file")
      or $self->throw("couldn't create Bio::SeqIO object");

  my $fh = $seqio->_fh or $self->throw("couldn't get filehandle from Bio::SeqIO object");
  my $offset    = tell($fh);
  $seqio->write_seq($seq);
  my $length    = tell($fh)-$offset;
  my $ids       = $self->seq_to_ids($seq);
  $self->_store_index($ids,$file,$offset,$length);

  $self->{flat_outfile_dirty}++;
}

sub close {
  my $self = shift;
  return unless $self->{flat_outfile_dirty};
  $self->write_config;
  delete $self->{flat_outfile_dirty};
  delete $self->{flat_cached_parsers}{$self->out_file};
}


sub _filenos {
  my $self = shift;
  return unless $self->{flat_flat_file_path};
  return keys %{$self->{flat_flat_file_path}};
}

# read the configuration file
sub _read_config {
  my $self   = shift;
  my $path = $self->_config_path;
  return unless -e $path;

  open (my $F,$path) or $self->throw("open error on $path: $!");
  my %config;
  while (<$F>) {
    chomp;
    my ($tag,@values) = split "\t";
    $config{$tag} = \@values;
  }
  CORE::close $F or $self->throw("close error on $path: $!");

  $config{index}[0] =~ m~(flat/1|BerkeleyDB/1)~
    or $self->throw("invalid configuration file $path: no index line");

  $self->indexing_scheme($1);

  if ($config{format}) {
    # handle LSID format
    if ($config{format}[0] =~ /^URN:LSID:open-bio\.org:(\w+)(?:\/(\w+))/) {
      $self->file_format($1);
      $self->alphabet($2);
    } else {  # compatibility with older versions
      $self->file_format($config{format}[0]);
    }
  }

  # set up primary namespace
  my $primary_namespace = $config{primary_namespace}[0]
    or $self->throw("invalid configuration file $path: no primary namespace defined");
  $self->primary_namespace($primary_namespace);

  # set up secondary namespaces (may be empty)
  $self->secondary_namespaces($config{secondary_namespaces});

  # get file paths and their normalization information
  my @normalized_files = grep {$_ ne ''} map {/^fileid_(\S+)/ && $1} keys %config;
  for my $nf (@normalized_files) {
    my ($file_path,$file_length) = @{$config{"fileid_${nf}"}};
    $self->add_flat_file($file_path,$file_length,$nf);
  }
  1;
}


sub _config_path {
  my $self = shift;
  $self->_catfile($self->_config_name);
}

sub _catfile {
  my $self = shift;
  my $component = shift;
  File::Spec->catfile($self->directory,$self->dbname,$component);
}

sub _config_name { CONFIG_FILE_NAME }

sub _path2fileno {
  my $self = shift;
  my $path = shift;
  return $self->add_flat_file($path)
    unless exists $self->{flat_flat_file_no}{$path};
  $self->{flat_flat_file_no}{$path};
}

sub _fileno2path {
  my $self = shift;
  my $fileno = shift;
  $self->{flat_flat_file_path}{$fileno};
}

sub _files {
  my $self = shift;
  my $paths = $self->{flat_flat_file_no};
  return keys %$paths;
}

=head2 fetch

  Title   : fetch
  Usage   : $index->fetch( $id )
  Function: Returns a Bio::Seq object from the index
  Example : $seq = $index->fetch( 'dJ67B12' )
  Returns : Bio::Seq object
  Args    : ID

Deprecated.  Use get_Seq_by_id instead.

=cut

sub fetch { shift->get_Seq_by_id(@_) }


=head2 To Be Implemented in Subclasses

The following methods MUST be implemented by subclasses.

=cut

# create real live Bio::Seq object
sub get_Seq_by_id {
  my $self = shift;
  my $id   = shift;
  $self->throw_not_implemented;
}


# fetch array of Bio::Seq objects
sub get_Seq_by_acc {
  my $self = shift;
  return $self->get_Seq_by_id(shift) if @_ == 1;
  my ($ns,$key) = @_;

  $self->throw_not_implemented;
}

sub fetch_raw {
  my ($self,$id,$namespace) = @_;
  $self->throw_not_implemented;
}

sub default_file_format {
  my $self = shift;
  $self->throw_not_implemented;
}

sub _store_index {
   my $self = shift;
   my ($ids,$file,$offset,$length) = @_;
   $self->throw_not_implemented;
}

=head2 May Be Overridden in Subclasses

The following methods MAY be overridden by subclasses.

=cut

sub default_primary_namespace {
  return "ACC";
}

sub default_secondary_namespaces {
  return;
}

sub seq_to_ids {
  my $self = shift;
  my $seq  = shift;
  my %ids;
  $ids{$self->primary_namespace} = $seq->accession_number;
  \%ids;
}

sub DESTROY {
  my $self = shift;
  $self->close;
}


1;