Diff of /Sprout/SproutLoad.pm

-revision 1.18, Wed Oct 12 03:17:58 2005 UTC
+revision 1.84, Thu May 17 23:44:51 2007 UTC
 Line 30
      $stats->Accumulate($spl->LoadFeatureData());
      print $stats->Show();
- This module makes use of the internal Sprout property C<_erdb>.
  It is worth noting that the FIG object does not need to be a real one. Any object
  that implements the FIG methods for data retrieval could be used. So, for example,
  this object could be used to copy data from one Sprout database to another, or
-Line 80
+Line 78
  =item subsysFile
  Either the name of the file containing the list of trusted subsystems or a reference
- to a list of subsystem names. If nothing is specified, all known subsystems will be
+ to a list of subsystem names. If nothing is specified, all NMPDR subsystems will be
- considered trusted. Only subsystem data related to the trusted subsystems is loaded.
+ considered trusted. (A subsystem is considered NMPDR if it has a file named C<NMPDR>
+ in its data directory.) Only subsystem data related to the NMPDR subsystems is loaded.
  =item options
-Line 94
+Line 93
  sub new {
      # Get the parameters.
      my ($class, $sprout, $fig, $genomeFile, $subsysFile, $options) = @_;
-     # Load the list of genomes into a hash.
+     # Create the genome hash.
-     my %genomes;
+     my %genomes = ();
+     # We only need it if load-only is NOT specified.
+     if (! $options->{loadOnly}) {
      if (! defined($genomeFile) || $genomeFile eq '') {
          # Here we want all the complete genomes and an access code of 1.
          my @genomeList = $fig->genomes(1);
-Line 119
+Line 120
                  # an omitted access code can be defaulted to 1.
                  for my $genomeLine (@genomeList) {
                      my ($genomeID, $accessCode) = split("\t", $genomeLine);
-                     if (undef $accessCode) {
+                         if (! defined($accessCode)) {
                          $accessCode = 1;
                      }
                      $genomes{$genomeID} = $accessCode;
-Line 129
+Line 130
              Confess("Invalid genome parameter ($type) in SproutLoad constructor.");
          }
      }
+     }
      # Load the list of trusted subsystems.
      my %subsystems = ();
+     # We only need it if load-only is NOT specified.
+     if (! $options->{loadOnly}) {
      if (! defined $subsysFile || $subsysFile eq '') {
-         # Here we want all the subsystems.
+             # Here we want all the usable subsystems. First we get the whole list.
-         %subsystems = map { $_ => 1 } $fig->all_subsystems();
+             my @subs = $fig->all_subsystems();
+             # Loop through, checking for the NMPDR file.
+             for my $sub (@subs) {
+                 if ($fig->nmpdr_subsystem($sub)) {
+                     $subsystems{$sub} = 1;
+                 }
+             }
      } else {
          my $type = ref $subsysFile;
          if ($type eq 'ARRAY') {
-Line 153
+Line 163
              Confess("Invalid subsystem parameter in SproutLoad constructor.");
          }
      }
+         # Go through the subsys hash again, creating the keyword list for each subsystem.
+         for my $subsystem (keys %subsystems) {
+             my $name = $subsystem;
+             $name =~ s/_/ /g;
+             my $classes = $fig->subsystem_classification($subsystem);
+             $name .= " " . join(" ", @{$classes});
+             $subsystems{$subsystem} = $name;
+         }
+     }
      # Get the data directory from the Sprout object.
      my ($directory) = $sprout->LoadInfo();
      # Create the Sprout load object.
-Line 162
+Line 181
                    subsystems => \%subsystems,
                    sprout => $sprout,
                    loadDirectory => $directory,
-                   erdb => $sprout->{_erdb},
+                   erdb => $sprout,
                    loaders => [],
                    options => $options
                   };
-Line 171
+Line 190
      return $retVal;
  }
+ =head3 LoadOnly
+ C<< my $flag = $spl->LoadOnly; >>
+ Return TRUE if we are in load-only mode, else FALSE.
+ =cut
+ sub LoadOnly {
+     my ($self) = @_;
+     return $self->{options}->{loadOnly};
+ }
+ =head3 PrimaryOnly
+ C<< my $flag = $spl->PrimaryOnly; >>
+ Return TRUE if only the main entity is to be loaded, else FALSE.
+ =cut
+ sub PrimaryOnly {
+     my ($self) = @_;
+     return $self->{options}->{primaryOnly};
+ }
  =head3 LoadGenomeData
  C<< my $stats = $spl->LoadGenomeData(); >>
-Line 198
+Line 243
  =back
- B<TO DO>
- Real quality vectors instead of C<unknown> for everything.
- GenomeGroup relation. (The original script took group information from the C<NMPDR> file
- in each genome's main directory, but no such file exists anywhere in my version of the
- data store.)
  =cut
  #: Return Type $%;
  sub LoadGenomeData {
-Line 216
+Line 253
      # Get the genome count.
      my $genomeHash = $self->{genomes};
      my $genomeCount = (keys %{$genomeHash});
-     Trace("Beginning genome data load.") if T(2);
      # Create load objects for each of the tables we're loading.
-     my $loadGenome = $self->_TableLoader('Genome', $genomeCount);
+     my $loadGenome = $self->_TableLoader('Genome');
-     my $loadHasContig = $self->_TableLoader('HasContig', $genomeCount * 300);
+     my $loadHasContig = $self->_TableLoader('HasContig', $self->PrimaryOnly);
-     my $loadContig = $self->_TableLoader('Contig', $genomeCount * 300);
+     my $loadContig = $self->_TableLoader('Contig', $self->PrimaryOnly);
-     my $loadIsMadeUpOf = $self->_TableLoader('IsMadeUpOf', $genomeCount * 60000);
+     my $loadIsMadeUpOf = $self->_TableLoader('IsMadeUpOf', $self->PrimaryOnly);
-     my $loadSequence = $self->_TableLoader('Sequence', $genomeCount * 60000);
+     my $loadSequence = $self->_TableLoader('Sequence', $self->PrimaryOnly);
+     if ($self->{options}->{loadOnly}) {
+         Trace("Loading from existing files.") if T(2);
+     } else {
+         Trace("Generating genome data.") if T(2);
      # Now we loop through the genomes, generating the data for each one.
      for my $genomeID (sort keys %{$genomeHash}) {
-         Trace("Loading data for genome $genomeID.") if T(3);
+             Trace("Generating data for genome $genomeID.") if T(3);
          $loadGenome->Add("genomeIn");
          # The access code comes in via the genome hash.
          my $accessCode = $genomeHash->{$genomeID};
-         # Get the genus, species, and strain from the scientific name. Note that we append
+             # Get the genus, species, and strain from the scientific name.
-         # the genome ID to the strain. In some cases this is the totality of the strain name.
          my ($genus, $species, @extraData) = split / /, $self->{fig}->genus_species($genomeID);
-         my $extra = join " ", @extraData, "[$genomeID]";
+             my $extra = join " ", @extraData;
          # Get the full taxonomy.
          my $taxonomy = $fig->taxonomy_of($genomeID);
+             # Get the version. If no version is specified, we default to the genome ID by itself.
+             my $version = $fig->genome_version($genomeID);
+             if (! defined($version)) {
+                 $version = $genomeID;
+             }
+             # Get the DNA size.
+             my $dnaSize = $fig->genome_szdna($genomeID);
+             # Open the NMPDR group file for this genome.
+             my $group;
+             if (open(TMP, "<$FIG_Config::organisms/$genomeID/NMPDR") &&
+                 defined($group = <TMP>)) {
+                 # Clean the line ending.
+                 chomp $group;
+             } else {
+                 # No group, so use the default.
+                 $group = $FIG_Config::otherGroup;
+             }
+             close TMP;
          # Output the genome record.
-         $loadGenome->Put($genomeID, $accessCode, $fig->is_complete($genomeID), $genus,
+             $loadGenome->Put($genomeID, $accessCode, $fig->is_complete($genomeID),
-                          $species, $extra, $taxonomy);
+                              $dnaSize, $genus, $group, $species, $extra, $version, $taxonomy);
          # Now we loop through each of the genome's contigs.
          my @contigs = $fig->all_contigs($genomeID);
          for my $contigID (@contigs) {
-Line 268
+Line 325
              }
          }
      }
+     }
      # Finish the loads.
      my $retVal = $self->_FinishAll();
      # Return the result.
-Line 308
+Line 366
      my $fig = $self->{fig};
      # Get the genome hash.
      my $genomeFilter = $self->{genomes};
-     my $genomeCount = (keys %{$genomeFilter});
+     # Set up an ID counter for the PCHs.
-     my $featureCount = $genomeCount * 4000;
+     my $pchID = 0;
      # Start the loads.
-     my $loadCoupling = $self->_TableLoader('Coupling', $featureCount * $genomeCount);
+     my $loadCoupling = $self->_TableLoader('Coupling');
-     my $loadIsEvidencedBy = $self->_TableLoader('IsEvidencedBy', $featureCount * 8000);
+     my $loadIsEvidencedBy = $self->_TableLoader('IsEvidencedBy', $self->PrimaryOnly);
-     my $loadPCH = $self->_TableLoader('PCH', $featureCount * 2000);
+     my $loadPCH = $self->_TableLoader('PCH', $self->PrimaryOnly);
-     my $loadParticipatesInCoupling = $self->_TableLoader('ParticipatesInCoupling', $featureCount * 2000);
+     my $loadParticipatesInCoupling = $self->_TableLoader('ParticipatesInCoupling', $self->PrimaryOnly);
-     my $loadUsesAsEvidence = $self->_TableLoader('UsesAsEvidence', $featureCount * 8000);
+     my $loadUsesAsEvidence = $self->_TableLoader('UsesAsEvidence', $self->PrimaryOnly);
-     Trace("Beginning coupling data load.") if T(2);
+     if ($self->{options}->{loadOnly}) {
+         Trace("Loading from existing files.") if T(2);
+     } else {
+         Trace("Generating coupling data.") if T(2);
      # Loop through the genomes found.
      for my $genome (sort keys %{$genomeFilter}) {
          Trace("Generating coupling data for $genome.") if T(3);
-Line 340
+Line 401
              for my $coupleData (@couplings) {
                  my ($peg2, $score) = @{$coupleData};
                  # Compute the coupling ID.
-                 my $coupleID = Sprout::CouplingID($peg1, $peg2);
+                     my $coupleID = $self->{erdb}->CouplingID($peg1, $peg2);
                  if (! exists $dupHash{$coupleID}) {
                      $loadCoupling->Add("couplingIn");
                      # Here we have a new coupling to store in the load files.
-Line 376
+Line 437
                          }
                      }
                      for my $evidenceID (keys %evidenceMap) {
+                             # Get the ID for this evidence.
+                             $pchID++;
                          # Create the evidence record.
                          my ($peg3, $peg4, $usage) = @{$evidenceMap{$evidenceID}};
-                         $loadPCH->Put($evidenceID, $usage);
+                             $loadPCH->Put($pchID, $usage);
                          # Connect it to the coupling.
-                         $loadIsEvidencedBy->Put($coupleID, $evidenceID);
+                             $loadIsEvidencedBy->Put($coupleID, $pchID);
                          # Connect it to the features.
-                         $loadUsesAsEvidence->Put($evidenceID, $peg3, 1);
+                             $loadUsesAsEvidence->Put($pchID, $peg3, 1);
-                         $loadUsesAsEvidence->Put($evidenceID, $peg4, 2);
+                             $loadUsesAsEvidence->Put($pchID, $peg4, 2);
+                         }
                      }
                  }
              }
-Line 410
+Line 474
      FeatureTranslation
      FeatureUpstream
      IsLocatedIn
+     HasFeature
+     HasRoleInSubsystem
+     FeatureEssential
+     FeatureVirulent
+     FeatureIEDB
  =over 4
-Line 424
+Line 493
  sub LoadFeatureData {
      # Get this object instance.
      my ($self) = @_;
-     # Get the FIG object.
+     # Get the FIG and Sprout objects.
      my $fig = $self->{fig};
-     # Find out if this is a limited run.
+     my $sprout = $self->{sprout};
-     my $limited = $self->{options}->{limitedFeatures};
      # Get the table of genome IDs.
      my $genomeHash = $self->{genomes};
-     my $genomeCount = (keys %{$genomeHash});
-     my $featureCount = $genomeCount * 4000;
      # Create load objects for each of the tables we're loading.
-     my $loadFeature = $self->_TableLoader('Feature', $featureCount);
+     my $loadFeature = $self->_TableLoader('Feature');
-     my $loadIsLocatedIn = $self->_TableLoader('IsLocatedIn', $featureCount);
+     my $loadIsLocatedIn = $self->_TableLoader('IsLocatedIn', $self->PrimaryOnly);
-     my $loadFeatureAlias = $self->_TableLoader('FeatureAlias', $featureCount * 6);
+     my $loadFeatureAlias = $self->_TableLoader('FeatureAlias');
-     my ($loadFeatureLink, $loadFeatureTranslation, $loadFeatureUpstream);
+     my $loadFeatureLink = $self->_TableLoader('FeatureLink');
-     if (! $limited) {
+     my $loadFeatureTranslation = $self->_TableLoader('FeatureTranslation');
-         $loadFeatureLink = $self->_TableLoader('FeatureLink', $featureCount * 10);
+     my $loadFeatureUpstream = $self->_TableLoader('FeatureUpstream');
-         $loadFeatureTranslation = $self->_TableLoader('FeatureTranslation', $featureCount);
+     my $loadHasFeature = $self->_TableLoader('HasFeature', $self->PrimaryOnly);
-         $loadFeatureUpstream = $self->_TableLoader('FeatureUpstream', $featureCount);
+     my $loadHasRoleInSubsystem = $self->_TableLoader('HasRoleInSubsystem', $self->PrimaryOnly);
-     }
+     my $loadFeatureEssential = $self->_TableLoader('FeatureEssential');
+     my $loadFeatureVirulent = $self->_TableLoader('FeatureVirulent');
+     my $loadFeatureIEDB = $self->_TableLoader('FeatureIEDB');
+     # Get the subsystem hash.
+     my $subHash = $self->{subsystems};
      # Get the maximum sequence size. We need this later for splitting up the
      # locations.
      my $chunkSize = $self->{sprout}->MaxSegment();
-     Trace("Beginning feature data load.") if T(2);
+     if ($self->{options}->{loadOnly}) {
+         Trace("Loading from existing files.") if T(2);
+     } else {
+         Trace("Generating feature data.") if T(2);
      # Now we loop through the genomes, generating the data for each one.
      for my $genomeID (sort keys %{$genomeHash}) {
          Trace("Loading features for genome $genomeID.") if T(3);
          $loadFeature->Add("genomeIn");
          # Get the feature list for this genome.
-         my $features = $fig->all_features_detailed($genomeID);
+             my $features = $fig->all_features_detailed_fast($genomeID);
+             # Sort and count the list.
+             my @featureTuples = sort { $a->[0] cmp $b->[0] } @{$features};
+             my $count = scalar @featureTuples;
+             my @fids = map { $_->[0] } @featureTuples;
+             Trace("$count features found for genome $genomeID.") if T(3);
+             # Get the attributes for this genome and put them in a hash by feature ID.
+             my $attributes = GetGenomeAttributes($fig, $genomeID, \@fids);
+             # Set up for our duplicate-feature check.
+             my $oldFeatureID = "";
          # Loop through the features.
-         for my $featureData (@{$features}) {
+             for my $featureTuple (@featureTuples) {
-             $loadFeature->Add("featureIn");
              # Split the tuple.
-             my ($featureID, $locations, undef, $type) = @{$featureData};
+                 my ($featureID, $locations, undef, $type, $minloc, $maxloc, $assignment, $user, $quality) = @{$featureTuple};
-             # Create the feature record.
+                 # Check for duplicates.
-             $loadFeature->Put($featureID, 1, $type);
+                 if ($featureID eq $oldFeatureID) {
+                     Trace("Duplicate feature $featureID found.") if T(1);
+                 } else {
+                     $oldFeatureID = $featureID;
+                     # Count this feature.
+                     $loadFeature->Add("featureIn");
+                     # Fix the quality. It is almost always a space, but some odd stuff might sneak through, and the
+                     # Sprout database requires a single character.
+                     if (! defined($quality) || $quality eq "") {
+                         $quality = " ";
+                     }
+                     # Begin building the keywords. We start with the genome ID, the
+                     # feature ID, the taxonomy, and the organism name.
+                     my @keywords = ($genomeID, $featureID, $fig->genus_species($genomeID),
+                                     $fig->taxonomy_of($genomeID));
              # Create the aliases.
              for my $alias ($fig->feature_aliases($featureID)) {
                  $loadFeatureAlias->Put($featureID, $alias);
+                         push @keywords, $alias;
              }
-             # The next stuff is for a full load only.
+                     Trace("Assignment for $featureID is: $assignment") if T(4);
-             if (! $limited) {
+                     # Break the assignment into words and shove it onto the
+                     # keyword list.
+                     push @keywords, split(/\s+/, $assignment);
+                     # Link this feature to the parent genome.
+                     $loadHasFeature->Put($genomeID, $featureID, $type);
                  # Get the links.
                  my @links = $fig->fid_links($featureID);
                  for my $link (@links) {
-Line 483
+Line 583
                          $loadFeatureUpstream->Put($featureID, $upstream);
                      }
                  }
-             }
+                     # Now we need to find the subsystems this feature participates in.
+                     # We also add the subsystems to the keyword list. Before we do that,
+                     # we must convert underscores to spaces and tack on the classifications.
+                     my @subsystems = $fig->peg_to_subsystems($featureID);
+                     for my $subsystem (@subsystems) {
+                         # Only proceed if we like this subsystem.
+                         if (exists $subHash->{$subsystem}) {
+                             # Store the has-role link.
+                             $loadHasRoleInSubsystem->Put($featureID, $subsystem, $genomeID, $type);
+                             # Save the subsystem's keyword data.
+                             my $subKeywords = $subHash->{$subsystem};
+                             push @keywords, split /\s+/, $subKeywords;
+                             # Now we need to get this feature's role in the subsystem.
+                             my $subObject = $fig->get_subsystem($subsystem);
+                             my @roleColumns = $subObject->get_peg_roles($featureID);
+                             my @allRoles = $subObject->get_roles();
+                             for my $col (@roleColumns) {
+                                 my $role = $allRoles[$col];
+                                 push @keywords, split /\s+/, $role;
+                                 push @keywords, $subObject->get_role_abbr($col);
+                             }
+                         }
+                     }
+                     # There are three special attributes computed from property
+                     # data that we build next. If the special attribute is non-empty,
+                     # its name will be added to the keyword list. First, we get all
+                     # the attributes for this feature. They will come back as
+                     # 4-tuples: [peg, name, value, URL]. We use a 3-tuple instead:
+                     # [name, value, value with URL]. (We don't need the PEG, since
+                     # we already know it.)
+                     my @attributes = map { [$_->[1], $_->[2], Tracer::CombineURL($_->[2], $_->[3])] }
+                                          @{$attributes->{$featureID}};
+                     # Now we process each of the special attributes.
+                     if (SpecialAttribute($featureID, \@attributes,
+, [0,2], '^(essential|potential_essential)$',
+                                          $loadFeatureEssential)) {
+                         push @keywords, 'essential';
+                         $loadFeature->Add('essential');
+                     }
+                     if (SpecialAttribute($featureID, \@attributes,
+, [2], '^virulen',
+                                          $loadFeatureVirulent)) {
+                         push @keywords, 'virulent';
+                         $loadFeature->Add('virulent');
+                     }
+                     if (SpecialAttribute($featureID, \@attributes,
+, [0,2], '^iedb_',
+                                          $loadFeatureIEDB)) {
+                         push @keywords, 'iedb';
+                         $loadFeature->Add('iedb');
+                     }
+                     # Now we need to bust up hyphenated words in the keyword
+                     # list. We keep them separate and put them at the end so
+                     # the original word order is available.
+                     my $keywordString = "";
+                     my $bustedString = "";
+                     for my $keyword (@keywords) {
+                         if (length $keyword >= 3) {
+                             $keywordString .= " $keyword";
+                             if ($keyword =~ /-/) {
+                                 my @words = split /-/, $keyword;
+                                 $bustedString .= join(" ", "", @words);
+                             }
+                         }
+                     }
+                     $keywordString .= $bustedString;
+                     # Get rid of annoying punctuation.
+                     $keywordString =~ s/[();]//g;
+                     # Clean the keyword list.
+                     my $cleanWords = $sprout->CleanKeywords($keywordString);
+                     Trace("Keyword string for $featureID: $cleanWords") if T(4);
+                     # Create the feature record.
+                     $loadFeature->Put($featureID, 1, $user, $quality, $type, $assignment, $cleanWords);
              # This part is the roughest. We need to relate the features to contig
              # locations, and the locations must be split so that none of them exceed
              # the maximum segment size. This simplifies the genes_in_region processing
-Line 512
+Line 684
              }
          }
      }
-     # Finish the loads.
-     my $retVal = $self->_FinishAll();
-     return $retVal;
- }
- =head3 LoadBBHData
- C<< my $stats = $spl->LoadBBHData(); >>
- Load the bidirectional best hit data from FIG into Sprout.
- Sprout does not store information on similarities. Instead, it has only the
- bi-directional best hits. Even so, the BBH table is one of the largest in
- the database.
- The following relations are loaded by this method.
-     IsBidirectionalBestHitOf
- =over 4
- =item RETURNS
- Returns a statistics object for the loads.
- =back
- =cut
- #: Return Type $%;
- sub LoadBBHData {
-     # Get this object instance.
-     my ($self) = @_;
-     # Get the FIG object.
-     my $fig = $self->{fig};
-     # Get the table of genome IDs.
-     my $genomeHash = $self->{genomes};
-     my $genomeCount = (keys %{$genomeHash});
-     my $featureCount = $genomeCount * 4000;
-     # Create load objects for each of the tables we're loading.
-     my $loadIsBidirectionalBestHitOf = $self->_TableLoader('IsBidirectionalBestHitOf',
-                                                            $featureCount * $genomeCount);
-     Trace("Beginning BBH load.") if T(2);
-     # Now we loop through the genomes, generating the data for each one.
-     for my $genomeID (sort keys %{$genomeHash}) {
-         $loadIsBidirectionalBestHitOf->Add("genomeIn");
-         Trace("Processing features for genome $genomeID.") if T(3);
-         # Get the feature list for this genome.
-         my $features = $fig->all_features_detailed($genomeID);
-         # Loop through the features.
-         for my $featureData (@{$features}) {
-             # Split the tuple.
-             my ($featureID, $locations, $aliases, $type) = @{$featureData};
-             # Get the bi-directional best hits.
-             my @bbhList = $fig->bbhs($featureID);
-             for my $bbhEntry (@bbhList) {
-                 # Get the target feature ID and the score.
-                 my ($targetID, $score) = @{$bbhEntry};
-                 # Check the target feature's genome.
-                 my $targetGenomeID = $fig->genome_of($targetID);
-                 # Only proceed if it's one of our genomes.
-                 if ($genomeHash->{$targetGenomeID}) {
-                     $loadIsBidirectionalBestHitOf->Put($featureID, $targetID, $targetGenomeID,
-                                                        $score);
-                 }
-             }
          }
      }
      # Finish the loads.
-Line 599
+Line 706
  The following relations are loaded by this method.
      Subsystem
+     SubsystemClass
      Role
+     RoleEC
      SSCell
      ContainsFeature
      IsGenomeOf
-Line 607
+Line 716
      OccursInSubsystem
      ParticipatesIn
      HasSSCell
-     Catalyzes
-     Reaction
      ConsistsOfRoles
      RoleSubset
      HasRoleSubset
      ConsistsOfGenomes
      GenomeSubset
      HasGenomeSubset
+     Catalyzes
+     Diagram
+     RoleOccursIn
  =over 4
-Line 624
+Line 734
  =back
- B<TO DO>
- Generate RoleName table?
  =cut
  #: Return Type $%;
  sub LoadSubsystemData {
-Line 641
+Line 747
      # Get the subsystem hash. This lists the subsystems we'll process.
      my $subsysHash = $self->{subsystems};
      my @subsysIDs = sort keys %{$subsysHash};
-     my $subsysCount = @subsysIDs;
+     # Get the map list.
-     my $genomeCount = (keys %{$genomeHash});
+     my @maps = $fig->all_maps;
-     my $featureCount = $genomeCount * 4000;
      # Create load objects for each of the tables we're loading.
-     my $loadSubsystem = $self->_TableLoader('Subsystem', $subsysCount);
+     my $loadDiagram = $self->_TableLoader('Diagram', $self->PrimaryOnly);
-     my $loadRole = $self->_TableLoader('Role', $featureCount * 6);
+     my $loadRoleOccursIn = $self->_TableLoader('RoleOccursIn', $self->PrimaryOnly);
-     my $loadSSCell = $self->_TableLoader('SSCell', $featureCount * $genomeCount);
+     my $loadSubsystem = $self->_TableLoader('Subsystem');
-     my $loadContainsFeature = $self->_TableLoader('ContainsFeature', $featureCount * $subsysCount);
+     my $loadRole = $self->_TableLoader('Role', $self->PrimaryOnly);
-     my $loadIsGenomeOf = $self->_TableLoader('IsGenomeOf', $featureCount * $genomeCount);
+     my $loadRoleEC = $self->_TableLoader('RoleEC', $self->PrimaryOnly);
-     my $loadIsRoleOf = $self->_TableLoader('IsRoleOf', $featureCount * $genomeCount);
+     my $loadCatalyzes = $self->_TableLoader('Catalyzes', $self->PrimaryOnly);
-     my $loadOccursInSubsystem = $self->_TableLoader('OccursInSubsystem', $featureCount * 6);
+     my $loadSSCell = $self->_TableLoader('SSCell', $self->PrimaryOnly);
-     my $loadParticipatesIn = $self->_TableLoader('ParticipatesIn', $subsysCount * $genomeCount);
+     my $loadContainsFeature = $self->_TableLoader('ContainsFeature', $self->PrimaryOnly);
-     my $loadHasSSCell = $self->_TableLoader('HasSSCell', $featureCount * $genomeCount);
+     my $loadIsGenomeOf = $self->_TableLoader('IsGenomeOf', $self->PrimaryOnly);
-     my $loadReaction = $self->_TableLoader('Reaction', $featureCount * $genomeCount);
+     my $loadIsRoleOf = $self->_TableLoader('IsRoleOf', $self->PrimaryOnly);
-     my $loadCatalyzes = $self->_TableLoader('Catalyzes', $featureCount * $genomeCount);
+     my $loadOccursInSubsystem = $self->_TableLoader('OccursInSubsystem', $self->PrimaryOnly);
-     my $loadRoleSubset = $self->_TableLoader('RoleSubset', $subsysCount * 50);
+     my $loadParticipatesIn = $self->_TableLoader('ParticipatesIn', $self->PrimaryOnly);
-     my $loadGenomeSubset = $self->_TableLoader('GenomeSubset', $subsysCount * 50);
+     my $loadHasSSCell = $self->_TableLoader('HasSSCell', $self->PrimaryOnly);
-     my $loadConsistsOfRoles = $self->_TableLoader('ConsistsOfRoles', $featureCount * $genomeCount);
+     my $loadRoleSubset = $self->_TableLoader('RoleSubset', $self->PrimaryOnly);
-     my $loadConsistsOfGenomes = $self->_TableLoader('ConsistsOfGenomes', $featureCount * $genomeCount);
+     my $loadGenomeSubset = $self->_TableLoader('GenomeSubset', $self->PrimaryOnly);
-     my $loadHasRoleSubset = $self->_TableLoader('HasRoleSubset', $subsysCount * 50);
+     my $loadConsistsOfRoles = $self->_TableLoader('ConsistsOfRoles', $self->PrimaryOnly);
-     my $loadHasGenomeSubset = $self->_TableLoader('HasGenomeSubset', $subsysCount * 50);
+     my $loadConsistsOfGenomes = $self->_TableLoader('ConsistsOfGenomes', $self->PrimaryOnly);
-     Trace("Beginning subsystem data load.") if T(2);
+     my $loadHasRoleSubset = $self->_TableLoader('HasRoleSubset', $self->PrimaryOnly);
-     # The reaction hash will contain a list of reactions for each role. When we're done,
+     my $loadHasGenomeSubset = $self->_TableLoader('HasGenomeSubset', $self->PrimaryOnly);
-     # a complicated sort and merge will be used to generate the Reaction and Catalyzes
+     my $loadSubsystemClass = $self->_TableLoader('SubsystemClass', $self->PrimaryOnly);
-     # tables.
+     if ($self->{options}->{loadOnly}) {
-     my %reactionsToRoles = ();
+         Trace("Loading from existing files.") if T(2);
+     } else {
+         Trace("Generating subsystem data.") if T(2);
+         # This hash will contain the role for each EC. When we're done, this
+         # information will be used to generate the Catalyzes table.
+         my %ecToRoles = ();
      # Loop through the subsystems. Our first task will be to create the
      # roles. We do this by looping through the subsystems and creating a
      # role hash. The hash tracks each role ID so that we don't create
-Line 675
+Line 785
      my ($genomeID, $roleID);
      my %roleData = ();
      for my $subsysID (@subsysIDs) {
-         Trace("Creating subsystem $subsysID.") if T(3);
-         $loadSubsystem->Add("subsystemIn");
          # Get the subsystem object.
          my $sub = $fig->get_subsystem($subsysID);
-         # Get its reaction hash.
+             # Only proceed if the subsystem has a spreadsheet.
-         my $reactionHash = $sub->get_reactions();
+             if (defined($sub) && ! $sub->{empty_ss}) {
+                 Trace("Creating subsystem $subsysID.") if T(3);
+                 $loadSubsystem->Add("subsystemIn");
          # Create the subsystem record.
          my $curator = $sub->get_curator();
          my $notes = $sub->get_notes();
          $loadSubsystem->Put($subsysID, $curator, $notes);
+                 # Now for the classification string. This comes back as a list
+                 # reference and we convert it to a space-delimited string.
+                 my $classList = $fig->subsystem_classification($subsysID);
+                 my $classString = join($FIG_Config::splitter, grep { $_ } @$classList);
+                 $loadSubsystemClass->Put($subsysID, $classString);
          # Connect it to its roles. Each role is a column in the subsystem spreadsheet.
          for (my $col = 0; defined($roleID = $sub->get_role($col)); $col++) {
              # Connect to this role.
-Line 697
+Line 812
                  # Add the role.
                  $loadRole->Put($roleID, $abbr);
                  $roleData{$roleID} = 1;
-                 # Add the role's reactions.
+                         # Check for an EC number.
-                 my $reactions = $reactionHash->{$roleID};
+                         if ($roleID =~ /\(EC ([^.]+\.[^.]+\.[^.]+\.[^)]+)\)\s*$/) {
-                 for my $reactionID (@{$reactions}) {
+                             my $ec = $1;
-                     if (! exists $reactionsToRoles{$reactionID}) {
+                             $loadRoleEC->Put($roleID, $ec);
-                         # Here the reaction is brand-new, so we create its reaction
+                             $ecToRoles{$ec} = $roleID;
-                         # record.
-                         $loadReaction->Put($reactionID, $fig->reversible($reactionID));
-                         # We also create a blank list for it in the reaction hash.
-                         $reactionsToRoles{$reactionID} = [];
-                     }
-                     # Add the role to the reaction's role list.
-                     push @{$reactionsToRoles{$reactionID}}, $roleID;
                  }
              }
          }
-Line 735
+Line 843
                  # part of the spreadsheet cell ID.
                  for (my $col = 0; defined($roleID = $sub->get_role($col)); $col++) {
                      # Get the features in the spreadsheet cell for this genome and role.
-                     my @pegs = $sub->get_pegs_from_cell($row, $col);
+                             my @pegs = grep { !$fig->is_deleted_fid($_) } $sub->get_pegs_from_cell($row, $col);
                      # Only proceed if features exist.
                      if (@pegs > 0) {
                          # Create the spreadsheet cell.
-Line 756
+Line 864
                  if ($pegCount > 0) {
                      Trace("$pegCount PEGs in $cellCount cells for $genomeID.") if T(3);
                      $loadParticipatesIn->Put($genomeID, $subsysID, $variantCode);
-                     # Partition the PEGs found into clusters.
-                     my @clusters = $fig->compute_clusters(\@pegsFound, $sub);
                      # Create a hash mapping PEG IDs to cluster numbers.
                      # We default to -1 for all of them.
                      my %clusterOf = map { $_ => -1 } @pegsFound;
+                             # Partition the PEGs found into clusters.
+                             my @clusters = $fig->compute_clusters([keys %clusterOf], $sub);
                      for (my $i = 0; $i <= $#clusters; $i++) {
                          my $subList = $clusters[$i];
                          for my $peg (@{$subList}) {
-Line 788
+Line 896
              # Connect the subset to the subsystem.
              $loadHasRoleSubset->Put($subsysID, $actualID);
              # Connect the subset to its roles.
-             my @roles = $sub->get_subset($subsetID);
+                     my @roles = $sub->get_subsetC_roles($subsetID);
              for my $roleID (@roles) {
                  $loadConsistsOfRoles->Put($actualID, $roleID);
              }
-Line 808
+Line 916
              }
          }
      }
-     # Before we leave, we must create the Catalyzes table. The data is all stored in
-     # "reactionToRoles" hash.
-     for my $reactionID (keys %reactionsToRoles) {
-         # Get this reaction's list of roles. We sort it so we can merge out duplicates.
-         my @roles = sort @{$reactionsToRoles{$reactionID}};
-         my $lastRole = "";
-         # Loop through the roles, creating catalyzation records.
-         for my $thisRole (@roles) {
-             if ($thisRole ne $lastRole) {
-                 $loadCatalyzes->Put($thisRole, $reactionID);
-             }
-         }
-     }
-     # Finish the load.
-     my $retVal = $self->_FinishAll();
-     return $retVal;
  }
+         # Now we loop through the diagrams. We need to create the diagram records
- =head3 LoadDiagramData
+         # and link each diagram to its roles. Note that only roles which occur
+         # in subsystems (and therefore appear in the %ecToRoles hash) are
- C<< my $stats = $spl->LoadDiagramData(); >>
+         # included.
+         for my $map (@maps) {
- Load the diagram data from FIG into Sprout.
- Diagrams are used to organize functional roles. The diagram shows the
- connections between chemicals that interact with a subsystem.
- The following relations are loaded by this method.
-     Diagram
-     RoleOccursIn
- =over 4
- =item RETURNS
- Returns a statistics object for the loads.
- =back
- =cut
- #: Return Type $%;
- sub LoadDiagramData {
-     # Get this object instance.
-     my ($self) = @_;
-     # Get the FIG object.
-     my $fig = $self->{fig};
-     # Get the map list.
-     my @maps = $fig->all_maps;
-     my $mapCount = @maps;
-     my $genomeCount = (keys %{$self->{genomes}});
-     my $featureCount = $genomeCount * 4000;
-     # Create load objects for each of the tables we're loading.
-     my $loadDiagram = $self->_TableLoader('Diagram', $mapCount);
-     my $loadRoleOccursIn = $self->_TableLoader('RoleOccursIn', $featureCount * 6);
-     Trace("Beginning diagram data load.") if T(2);
-     # Loop through the diagrams.
-     for my $map ($fig->all_maps) {
          Trace("Loading diagram $map.") if T(3);
          # Get the diagram's descriptive name.
          my $name = $fig->map_name($map);
-Line 874
+Line 930
          # A hash is used to prevent duplicates.
          my %roleHash = ();
          for my $role ($fig->map_to_ecs($map)) {
-             if (! $roleHash{$role}) {
+                 if (exists $ecToRoles{$role} && ! $roleHash{$role}) {
-                 $loadRoleOccursIn->Put($role, $map);
+                     $loadRoleOccursIn->Put($ecToRoles{$role}, $map);
                  $roleHash{$role} = 1;
              }
          }
      }
+         # Before we leave, we must create the Catalyzes table. We start with the reactions,
+         # then use the "ecToRoles" table to convert EC numbers to role IDs.
+         my @reactions = $fig->all_reactions();
+         for my $reactionID (@reactions) {
+             # Get this reaction's list of roles. The results will be EC numbers.
+             my @roles = $fig->catalyzed_by($reactionID);
+             # Loop through the roles, creating catalyzation records.
+             for my $thisRole (@roles) {
+                 if (exists $ecToRoles{$thisRole}) {
+                     $loadCatalyzes->Put($ecToRoles{$thisRole}, $reactionID);
+                 }
+             }
+         }
+     }
      # Finish the load.
      my $retVal = $self->_FinishAll();
      return $retVal;
-Line 921
+Line 991
      my $fig = $self->{fig};
      # Get the genome hash.
      my $genomeHash = $self->{genomes};
-     my $genomeCount = (keys %{$genomeHash});
      # Create load objects for each of the tables we're loading.
-     my $loadProperty = $self->_TableLoader('Property', $genomeCount * 1500);
+     my $loadProperty = $self->_TableLoader('Property');
-     my $loadHasProperty = $self->_TableLoader('HasProperty', $genomeCount * 1500);
+     my $loadHasProperty = $self->_TableLoader('HasProperty', $self->PrimaryOnly);
-     Trace("Beginning property data load.") if T(2);
+     if ($self->{options}->{loadOnly}) {
+         Trace("Loading from existing files.") if T(2);
+     } else {
+         Trace("Generating property data.") if T(2);
      # Create a hash for storing property IDs.
      my %propertyKeys = ();
      my $nextID = 1;
+         # Get the attributes we intend to store in the property table.
+         my @propKeys = $fig->get_group_keys("NMPDR");
      # Loop through the genomes.
-     for my $genomeID (keys %{$genomeHash}) {
+         for my $genomeID (sort keys %{$genomeHash}) {
          $loadProperty->Add("genomeIn");
-         # Get the genome's features. The feature ID is the first field in the
+             Trace("Generating properties for $genomeID.") if T(3);
-         # tuples returned by "all_features_detailed". We use "all_features_detailed"
+             # Initialize a counter.
-         # rather than "all_features" because we want all features regardless of type.
+             my $propertyCount = 0;
-         my @features = map { $_->[0] } @{$fig->all_features_detailed($genomeID)};
+             # Get the properties for this genome's features.
-         # Loop through the features, creating HasProperty records.
+             my @attributes = $fig->get_attributes("fig|$genomeID%", \@propKeys);
-         for my $fid (@features) {
+             Trace("Property list built for $genomeID.") if T(3);
-             $loadProperty->Add("featureIn");
+             # Loop through the results, creating HasProperty records.
-             # Get all attributes for this feature. We do this one feature at a time
+             for my $attributeData (@attributes) {
-             # to insure we do not get any genome attributes.
+                 # Pull apart the attribute tuple.
-             my @attributeList = $fig->get_attributes($fid, '', '', '');
+                 my ($fid, $key, $value, $url) = @{$attributeData};
-             # Loop through the attributes.
-             for my $tuple (@attributeList) {
-                 # Get this attribute value's data. Note that we throw away the FID,
-                 # since it will always be the same as the value if "$fid".
-                 my (undef, $key, $value, $url) = @{$tuple};
                  # Concatenate the key and value and check the "propertyKeys" hash to
                  # see if we already have an ID for it. We use a tab for the separator
                  # character.
-Line 964
+Line 1033
                  # Create the HasProperty entry for this feature/property association.
                  $loadHasProperty->Put($fid, $propertyID, $url);
              }
+             # Update the statistics.
+             Trace("$propertyCount attributes processed.") if T(3);
+             $loadHasProperty->Add("propertiesIn", $propertyCount);
          }
      }
      # Finish the load.
-Line 1005
+Line 1077
      my $fig = $self->{fig};
      # Get the genome hash.
      my $genomeHash = $self->{genomes};
-     my $genomeCount = (keys %{$genomeHash});
      # Create load objects for each of the tables we're loading.
-     my $loadAnnotation = $self->_TableLoader('Annotation', $genomeCount * 4000);
+     my $loadAnnotation = $self->_TableLoader('Annotation');
-     my $loadIsTargetOfAnnotation = $self->_TableLoader('IsTargetOfAnnotation', $genomeCount * 4000);
+     my $loadIsTargetOfAnnotation = $self->_TableLoader('IsTargetOfAnnotation', $self->PrimaryOnly);
-     my $loadSproutUser = $self->_TableLoader('SproutUser', 100);
+     my $loadSproutUser = $self->_TableLoader('SproutUser', $self->PrimaryOnly);
-     my $loadUserAccess = $self->_TableLoader('UserAccess', 1000);
+     my $loadUserAccess = $self->_TableLoader('UserAccess', $self->PrimaryOnly);
-     my $loadMadeAnnotation = $self->_TableLoader('MadeAnnotation', $genomeCount * 4000);
+     my $loadMadeAnnotation = $self->_TableLoader('MadeAnnotation', $self->PrimaryOnly);
-     Trace("Beginning annotation data load.") if T(2);
+     if ($self->{options}->{loadOnly}) {
+         Trace("Loading from existing files.") if T(2);
+     } else {
+         Trace("Generating annotation data.") if T(2);
      # Create a hash of user names. We'll use this to prevent us from generating duplicate
      # user records.
      my %users = ( FIG => 1, master => 1 );
-Line 1026
+Line 1100
      # Loop through the genomes.
      for my $genomeID (sort keys %{$genomeHash}) {
          Trace("Processing $genomeID.") if T(3);
-         # Get the genome's PEGs.
-         my @pegs = $fig->pegs_of($genomeID);
-         for my $peg (@pegs) {
-             Trace("Processing $peg.") if T(4);
              # Create a hash of timestamps. We use this to prevent duplicate time stamps
              # from showing up for a single PEG's annotations.
              my %seenTimestamps = ();
-             # Check for a functional assignment.
+             # Get the genome's annotations.
-             my $func = $fig->function_of($peg);
+             my @annotations = $fig->read_all_annotations($genomeID);
-             if ($func) {
+             Trace("Processing annotations.") if T(2);
-                 # If this is NOT a hypothetical assignment, we create an
+             for my $tuple (@annotations) {
-                 # assignment annotation for it.
+                 # Get the annotation tuple.
-                 if (! FIG::hypo($peg)) {
+                 my ($peg, $timestamp, $user, $text) = @{$tuple};
-                     # Note that we double the slashes so that what goes into the database is
-                     # a new-line escape sequence rather than an actual new-line.
-                     $loadAnnotation->Put("$peg:$time", $time, "FIG\\nSet function to\\n$func");
-                     $loadIsTargetOfAnnotation->Put($peg, "$peg:$time");
-                     $loadMadeAnnotation->Put("FIG", "$peg:$time");
-                     # Denote we've seen this timestamp.
-                     $seenTimestamps{$time} = 1;
-                 }
-             }
-             # Now loop through the real annotations.
-             for my $tuple ($fig->feature_annotations($peg, "raw")) {
-                 my ($fid, $timestamp, $user, $text) = @{$tuple};
                  # Here we fix up the annotation text. "\r" is removed,
-                 # and "\t" and "\n" are escaped. Note we use the "s"
+                 # and "\t" and "\n" are escaped. Note we use the "gs"
                  # modifier so that new-lines inside the text do not
                  # stop the substitution search.
                  $text =~ s/\r//gs;
-Line 1065
+Line 1123
                      # Here it's a number. We need to insure the one we use to form
                      # the key is unique.
                      my $keyStamp = $timestamp;
-                     while ($seenTimestamps{$keyStamp}) {
+                     while ($seenTimestamps{"$peg:$keyStamp"}) {
                          $keyStamp++;
                      }
-                     $seenTimestamps{$keyStamp} = 1;
                      my $annotationID = "$peg:$keyStamp";
+                     $seenTimestamps{$annotationID} = 1;
                      # Insure the user exists.
                      if (! $users{$user}) {
                          $loadSproutUser->Put($user, "SEED user");
-Line 1127
+Line 1185
      my $fig = $self->{fig};
      # Get the genome hash.
      my $genomeHash = $self->{genomes};
-     my $genomeCount = (keys %{$genomeHash});
      # Create load objects for each of the tables we're loading.
-     my $loadComesFrom = $self->_TableLoader('ComesFrom', $genomeCount * 4);
+     my $loadComesFrom = $self->_TableLoader('ComesFrom', $self->PrimaryOnly);
-     my $loadSource = $self->_TableLoader('Source', $genomeCount * 4);
+     my $loadSource = $self->_TableLoader('Source');
-     my $loadSourceURL = $self->_TableLoader('SourceURL', $genomeCount * 8);
+     my $loadSourceURL = $self->_TableLoader('SourceURL');
-     Trace("Beginning source data load.") if T(2);
+     if ($self->{options}->{loadOnly}) {
+         Trace("Loading from existing files.") if T(2);
+     } else {
+         Trace("Generating annotation data.") if T(2);
      # Create hashes to collect the Source information.
      my %sourceURL = ();
      my %sourceDesc = ();
-Line 1162
+Line 1222
      for my $sourceID (keys %sourceDesc) {
          $loadSource->Put($sourceID, $sourceDesc{$sourceID});
      }
+     }
      # Finish the load.
      my $retVal = $self->_FinishAll();
      return $retVal;
-Line 1201
+Line 1262
      my $fig = $self->{fig};
      # Get the genome hash.
      my $genomeHash = $self->{genomes};
-     my $genomeCount = (keys %{$genomeHash});
      # Convert the genome hash. We'll get the genus and species for each genome and make
      # it the key.
      my %speciesHash = map { $fig->genus_species($_) => $_ } (keys %{$genomeHash});
      # Create load objects for each of the tables we're loading.
-     my $loadExternalAliasFunc = $self->_TableLoader('ExternalAliasFunc', $genomeCount * 4000);
+     my $loadExternalAliasFunc = $self->_TableLoader('ExternalAliasFunc');
-     my $loadExternalAliasOrg = $self->_TableLoader('ExternalAliasOrg', $genomeCount * 4000);
+     my $loadExternalAliasOrg = $self->_TableLoader('ExternalAliasOrg');
-     Trace("Beginning external data load.") if T(2);
+     if ($self->{options}->{loadOnly}) {
+         Trace("Loading from existing files.") if T(2);
+     } else {
+         Trace("Generating external data.") if T(2);
      # We loop through the files one at a time. First, the organism file.
-     Open(\*ORGS, "<$FIG_Config::global/ext_org.table");
+         Open(\*ORGS, "sort +0 -1 -u -t\"\t\" $FIG_Config::global/ext_org.table |");
      my $orgLine;
      while (defined($orgLine = <ORGS>)) {
          # Clean the input line.
-Line 1222
+Line 1285
      close ORGS;
      # Now the function file.
      my $funcLine;
-     Open(\*FUNCS, "<$FIG_Config::global/ext_func.table");
+         Open(\*FUNCS, "sort +0 -1 -u -t\"\t\" $FIG_Config::global/ext_func.table |");
      while (defined($funcLine = <FUNCS>)) {
          # Clean the line ending.
          chomp $funcLine;
-Line 1238
+Line 1301
              $loadExternalAliasFunc->Put(@funcFields[0,1]);
          }
      }
+     }
      # Finish the load.
      my $retVal = $self->_FinishAll();
      return $retVal;
-Line 1254
+Line 1318
  The following relations are loaded by this method.
+     Reaction
      ReactionURL
      Compound
      CompoundName
-Line 1277
+Line 1342
      my ($self) = @_;
      # Get the FIG object.
      my $fig = $self->{fig};
-     # Get the genome hash.
-     my $genomeHash = $self->{genomes};
-     my $genomeCount = (keys %{$genomeHash});
      # Create load objects for each of the tables we're loading.
-     my $loadReactionURL = $self->_TableLoader('ReactionURL', $genomeCount * 4000);
+     my $loadReaction = $self->_TableLoader('Reaction');
-     my $loadCompound = $self->_TableLoader('Compound', $genomeCount * 4000);
+     my $loadReactionURL = $self->_TableLoader('ReactionURL', $self->PrimaryOnly);
-     my $loadCompoundName = $self->_TableLoader('CompoundName', $genomeCount * 8000);
+     my $loadCompound = $self->_TableLoader('Compound', $self->PrimaryOnly);
-     my $loadCompoundCAS = $self->_TableLoader('CompoundCAS', $genomeCount * 4000);
+     my $loadCompoundName = $self->_TableLoader('CompoundName', $self->PrimaryOnly);
-     my $loadIsAComponentOf = $self->_TableLoader('IsAComponentOf', $genomeCount * 12000);
+     my $loadCompoundCAS = $self->_TableLoader('CompoundCAS', $self->PrimaryOnly);
-     Trace("Beginning reaction/compound data load.") if T(2);
+     my $loadIsAComponentOf = $self->_TableLoader('IsAComponentOf', $self->PrimaryOnly);
-     # Create a hash to remember the compounds we've generated in the compound table.
+     if ($self->{options}->{loadOnly}) {
-     my %compoundHash = ();
+         Trace("Loading from existing files.") if T(2);
-     # Loop through the reactions.
+     } else {
+         Trace("Generating annotation data.") if T(2);
+         # First we create the compounds.
+         my @compounds = $fig->all_compounds();
+         for my $cid (@compounds) {
+             # Check for names.
+             my @names = $fig->names_of_compound($cid);
+             # Each name will be given a priority number, starting with 1.
+             my $prio = 1;
+             for my $name (@names) {
+                 $loadCompoundName->Put($cid, $name, $prio++);
+             }
+             # Create the main compound record. Note that the first name
+             # becomes the label.
+             my $label = (@names > 0 ? $names[0] : $cid);
+             $loadCompound->Put($cid, $label);
+             # Check for a CAS ID.
+             my $cas = $fig->cas($cid);
+             if ($cas) {
+                 $loadCompoundCAS->Put($cid, $cas);
+             }
+         }
+         # All the compounds are set up, so we need to loop through the reactions next. First,
+         # we initialize the discriminator index. This is a single integer used to insure
+         # duplicate elements in a reaction are not accidentally collapsed.
+         my $discrim = 0;
      my @reactions = $fig->all_reactions();
      for my $reactionID (@reactions) {
+             # Create the reaction record.
+             $loadReaction->Put($reactionID, $fig->reversible($reactionID));
          # Compute the reaction's URL.
          my $url = HTML::reaction_link($reactionID);
          # Put it in the ReactionURL table.
-Line 1300
+Line 1389
          # substrates first and then products.
          for my $product (0, 1) {
              # Get the compounds of the current type for the current reaction. FIG will
-             # give us 3-tuples: [ID, Stoichometry, main-flag]. At this time we do not
+                 # give us 3-tuples: [ID, stoichiometry, main-flag]. At this time we do not
              # have location data in SEED, so it defaults to the empty string.
              my @compounds = $fig->reaction2comp($reactionID, $product);
              for my $compData (@compounds) {
                  # Extract the compound data from the current tuple.
                  my ($cid, $stoich, $main) = @{$compData};
                  # Link the compound to the reaction.
-                 $loadIsAComponentOf->Put($cid, $reactionID, "", $main, $product, $stoich);
+                     $loadIsAComponentOf->Put($cid, $reactionID, $discrim++, "", $main,
-                 # If this is a new compound, we need to create its table entries.
+                                              $product, $stoich);
-                 if (! exists $compoundHash{$cid}) {
-                     $compoundHash{$cid} = 1;
-                     # Create the main compound record and denote we've done it.
-                     $loadCompound->Put($cid);
-                     # Check for a CAS ID.
-                     my $cas = $fig->cas($cid);
-                     if ($cas) {
-                         $loadCompoundCAS->Put($cid, $cas);
-                     }
-                     # Check for names.
-                     my @names = $fig->names_of_compound($cid);
-                     # Each name will be given a priority number, starting with 1.
-                     my $prio = 0;
-                     for my $name (@names) {
-                         $loadCompoundName->Put($cid, $name, $prio++);
-                     }
                  }
              }
          }
-Line 1344
+Line 1417
      GenomeGroups
- There is no direct support for genome groups in FIG, so we access the SEED
+ Currently, we do not use groups. We used to use them for NMPDR groups,
+ butThere is no direct support for genome groups in FIG, so we access the SEED
  files directly.
  =over 4
-Line 1364
+Line 1438
      my $fig = $self->{fig};
      # Get the genome hash.
      my $genomeHash = $self->{genomes};
-     my $genomeCount = (keys %{$genomeHash});
      # Create a load object for the table we're loading.
-     my $loadGenomeGroups = $self->_TableLoader('GenomeGroups', $genomeCount * 4);
+     my $loadGenomeGroups = $self->_TableLoader('GenomeGroups');
-     Trace("Beginning group data load.") if T(2);
+     if ($self->{options}->{loadOnly}) {
+         Trace("Loading from existing files.") if T(2);
+     } else {
+         Trace("Generating group data.") if T(2);
+         # Currently there are no groups.
+     }
+     # Finish the load.
+     my $retVal = $self->_FinishAll();
+     return $retVal;
+ }
+ =head3 LoadSynonymData
+ C<< my $stats = $spl->LoadSynonymData(); >>
+ Load the synonym groups into Sprout.
+ The following relations are loaded by this method.
+     SynonymGroup
+     IsSynonymGroupFor
+ The source information for these relations is taken from the C<maps_to_id> method
+ of the B<FIG> object. Unfortunately, to make this work, we need to use direct
+ SQL against the FIG database.
+ =over 4
+ =item RETURNS
+ Returns a statistics object for the loads.
+ =back
+ =cut
+ #: Return Type $%;
+ sub LoadSynonymData {
+     # Get this object instance.
+     my ($self) = @_;
+     # Get the FIG object.
+     my $fig = $self->{fig};
+     # Get the genome hash.
+     my $genomeHash = $self->{genomes};
+     # Create a load object for the table we're loading.
+     my $loadSynonymGroup = $self->_TableLoader('SynonymGroup');
+     my $loadIsSynonymGroupFor = $self->_TableLoader('IsSynonymGroupFor');
+     if ($self->{options}->{loadOnly}) {
+         Trace("Loading from existing files.") if T(2);
+     } else {
+         Trace("Generating synonym group data.") if T(2);
+         # Get the database handle.
+         my $dbh = $fig->db_handle();
+         # Ask for the synonyms.
+         my $sth = $dbh->prepare_command("SELECT maps_to, syn_id FROM peg_synonyms ORDER BY maps_to");
+         my $result = $sth->execute();
+         if (! defined($result)) {
+             Confess("Database error in Synonym load: " . $sth->errstr());
+         } else {
+             # Remember the current synonym.
+             my $current_syn = "";
+             # Count the features.
+             my $featureCount = 0;
+             # Loop through the synonym/peg pairs.
+             while (my @row = $sth->fetchrow()) {
+                 # Get the synonym ID and feature ID.
+                 my ($syn_id, $peg) = @row;
+                 # Insure it's for one of our genomes.
+                 my $genomeID = FIG::genome_of($peg);
+                 if (exists $genomeHash->{$genomeID}) {
+                     # Verify the synonym.
+                     if ($syn_id ne $current_syn) {
+                         # It's new, so put it in the group table.
+                         $loadSynonymGroup->Put($syn_id);
+                         $current_syn = $syn_id;
+                     }
+                     # Connect the synonym to the peg.
+                     $loadIsSynonymGroupFor->Put($syn_id, $peg);
+                     # Count this feature.
+                     $featureCount++;
+                     if ($featureCount % 1000 == 0) {
+                         Trace("$featureCount features processed.") if T(3);
+                     }
+                 }
+             }
+         }
+     }
+     # Finish the load.
+     my $retVal = $self->_FinishAll();
+     return $retVal;
+ }
+ =head3 LoadFamilyData
+ C<< my $stats = $spl->LoadFamilyData(); >>
+ Load the protein families into Sprout.
+ The following relations are loaded by this method.
+     Family
+     IsFamilyForFeature
+ The source information for these relations is taken from the C<families_for_protein>,
+ C<family_function>, and C<sz_family> methods of the B<FIG> object.
+ =over 4
+ =item RETURNS
+ Returns a statistics object for the loads.
+ =back
+ =cut
+ #: Return Type $%;
+ sub LoadFamilyData {
+     # Get this object instance.
+     my ($self) = @_;
+     # Get the FIG object.
+     my $fig = $self->{fig};
+     # Get the genome hash.
+     my $genomeHash = $self->{genomes};
+     # Create load objects for the tables we're loading.
+     my $loadFamily = $self->_TableLoader('Family');
+     my $loadIsFamilyForFeature = $self->_TableLoader('IsFamilyForFeature');
+     if ($self->{options}->{loadOnly}) {
+         Trace("Loading from existing files.") if T(2);
+     } else {
+         Trace("Generating family data.") if T(2);
+         # Create a hash for the family IDs.
+         my %familyHash = ();
      # Loop through the genomes.
-     my $line;
+         for my $genomeID (sort keys %{$genomeHash}) {
-     for my $genomeID (keys %{$genomeHash}) {
+             Trace("Processing features for $genomeID.") if T(2);
-         Trace("Processing $genomeID.") if T(3);
+             # Loop through this genome's PEGs.
-         # Open the NMPDR group file for this genome.
+             for my $fid ($fig->all_features($genomeID, "peg")) {
-         if (open(TMP, "<$FIG_Config::organisms/$genomeID/NMPDR") &&
+                 $loadIsFamilyForFeature->Add("features", 1);
-             defined($line = <TMP>)) {
+                 # Get this feature's families.
-             # Clean the line ending.
+                 my @families = $fig->families_for_protein($fid);
-             chomp $line;
+                 # Loop through the families, connecting them to the feature.
-             # Add the group to the table. Note that there can only be one group
+                 for my $family (@families) {
-             # per genome.
+                     $loadIsFamilyForFeature->Put($family, $fid);
-             $loadGenomeGroups->Put($genomeID, $line);
+                     # If this is a new family, create a record for it.
+                     if (! exists $familyHash{$family}) {
+                         $familyHash{$family} = 1;
+                         $loadFamily->Add("families", 1);
+                         my $size = $fig->sz_family($family);
+                         my $func = $fig->family_function($family);
+                         $loadFamily->Put($family, $size, $func);
+                     }
+                 }
+             }
          }
-         close TMP;
      }
      # Finish the load.
      my $retVal = $self->_FinishAll();
      return $retVal;
  }
+ =head3 LoadDrugData
+ C<< my $stats = $spl->LoadDrugData(); >>
+ Load the drug target data into Sprout.
+ The following relations are loaded by this method.
+     PDB
+     DocksWith
+     IsProteinForFeature
+     Ligand
+ The source information for these relations is taken from attributes. The
+ C<PDB> attribute links a PDB to a feature, and is used to build B<IsProteinForFeature>.
+ The C<zinc_name> attribute describes the ligands. The C<docking_results>
+ attribute contains the information for the B<DocksWith> relationship. It is
+ expected that additional attributes and tables will be added in the future.
+ =over 4
+ =item RETURNS
+ Returns a statistics object for the loads.
+ =back
+ =cut
+ #: Return Type $%;
+ sub LoadDrugData {
+     # Get this object instance.
+     my ($self) = @_;
+     # Get the FIG object.
+     my $fig = $self->{fig};
+     # Get the genome hash.
+     my $genomeHash = $self->{genomes};
+     # Create load objects for the tables we're loading.
+     my $loadPDB = $self->_TableLoader('PDB');
+     my $loadLigand = $self->_TableLoader('Ligand');
+     my $loadIsProteinForFeature = $self->_TableLoader('IsProteinForFeature');
+     my $loadDocksWith = $self->_TableLoader('DocksWith');
+     if ($self->{options}->{loadOnly}) {
+         Trace("Loading from existing files.") if T(2);
+     } else {
+         Trace("Generating drug target data.") if T(2);
+         # First comes the "DocksWith" relationship. This will give us a list of PDBs.
+         # We can also encounter PDBs when we process "IsProteinForFeature". To manage
+         # this process, PDB information is collected in a hash table and then
+         # unspooled after both relationships are created.
+         my %pdbHash = ();
+         Trace("Generating docking data.") if T(2);
+         # Get all the docking data. This may cause problems if there are too many PDBs,
+         # at which point we'll need another algorithm. The indicator that this is
+         # happening will be a timeout error in the next statement.
+         my @dockData = $fig->query_attributes('$key = ? AND $value < ?',
+                                               ['docking_results', $FIG_Config::dockLimit]);
+         Trace(scalar(@dockData) . " rows of docking data found.") if T(3);
+         for my $dockData (@dockData) {
+             # Get the docking data components.
+             my ($pdbID, $docking_key, @valueData) = @{$dockData};
+             # Fix the PDB ID. It's supposed to be lower-case, but this does not always happen.
+             $pdbID = lc $pdbID;
+             # Strip off the object type.
+             $pdbID =~ s/pdb://;
+             # Extract the ZINC ID from the docking key. Note that there are two possible
+             # formats.
+             my (undef, $zinc_id) = $docking_key =~ /^docking_results::(ZINC)?(\d+)$/;
+             if (! $zinc_id) {
+                 Trace("Invalid docking result key $docking_key for $pdbID.") if T(0);
+                 $loadDocksWith->Add("errors");
+             } else {
+                 # Get the pieces of the value and parse the energy.
+                 # Note that we don't care about the rank, since
+                 # we can sort on the energy level itself in our database.
+                 my ($energy, $tool, $type) = @valueData;
+                 my ($rank, $total, $vanderwaals, $electrostatic) = split /\s*;\s*/, $energy;
+                 # Ignore predicted results.
+                 if ($type ne "Predicted") {
+                     # Count this docking result.
+                     if (! exists $pdbHash{$pdbID}) {
+                         $pdbHash{$pdbID} = 1;
+                     } else {
+                         $pdbHash{$pdbID}++;
+                     }
+                     # Write the result to the output.
+                     $loadDocksWith->Put($pdbID, $zinc_id, $electrostatic, $type, $tool,
+                                         $total, $vanderwaals);
+                 }
+             }
+         }
+         Trace("Connecting features.") if T(2);
+         # Loop through the genomes.
+         for my $genome (sort keys %{$genomeHash}) {
+             Trace("Generating PDBs for $genome.") if T(3);
+             # Get all of the PDBs that BLAST against this genome's features.
+             my @attributeData = $fig->get_attributes("fig|$genome%", 'PDB::%');
+             for my $pdbData (@attributeData) {
+                 # The PDB ID is coded as a subkey.
+                 if ($pdbData->[1] !~ /PDB::(.+)/i) {
+                     Trace("Invalid PDB ID \"$pdbData->[1]\" in attribute table.") if T(0);
+                     $loadPDB->Add("errors");
+                 } else {
+                     my $pdbID = $1;
+                     # Insure the PDB is in the hash.
+                     if (! exists $pdbHash{$pdbID}) {
+                         $pdbHash{$pdbID} = 0;
+                     }
+                     # The score and locations are coded in the attribute value.
+                     if ($pdbData->[2] !~ /^([^;]+)(.*)$/) {
+                         Trace("Invalid PDB data for $pdbID and feature $pdbData->[0].") if T(0);
+                         $loadIsProteinForFeature->Add("errors");
+                     } else {
+                         my ($score, $locData) = ($1,$2);
+                         # The location data may not be present, so we have to start with some
+                         # defaults and then check.
+                         my ($start, $end) = (1, 0);
+                         if ($locData) {
+                             $locData =~ /(\d+)-(\d+)/;
+                             $start = $1;
+                             $end = $2;
+                         }
+                         # If we still don't have the end location, compute it from
+                         # the feature length.
+                         if (! $end) {
+                             # Most features have one location, but we do a list iteration
+                             # just in case.
+                             my @locations = $fig->feature_location($pdbData->[0]);
+                             $end = 0;
+                             for my $loc (@locations) {
+                                 my $locObject = BasicLocation->new($loc);
+                                 $end += $locObject->Length;
+                             }
+                         }
+                         # Decode the score.
+                         my $realScore = FIGRules::DecodeScore($score);
+                         # Connect the PDB to the feature.
+                         $loadIsProteinForFeature->Put($pdbData->[0], $pdbID, $start, $realScore, $end);
+                     }
+                 }
+             }
+         }
+         # We've got all our PDBs now, so we unspool them from the hash.
+         Trace("Generating PDBs. " . scalar(keys %pdbHash) . " found.") if T(2);
+         my $count = 0;
+         for my $pdbID (sort keys %pdbHash) {
+             $loadPDB->Put($pdbID, $pdbHash{$pdbID});
+             $count++;
+             Trace("$count PDBs processed.") if T(3) && ($count % 500 == 0);
+         }
+         # Finally we create the ligand table. This information can be found in the
+         # zinc_name attribute.
+         Trace("Loading ligands.") if T(2);
+         # The ligand list is huge, so we have to get it in pieces. We also have to check for duplicates.
+         my $last_zinc_id = "";
+         my $zinc_id = "";
+         my $done = 0;
+         while (! $done) {
+             # Get the next 10000 ligands. We insist that the object ID is greater than
+             # the last ID we processed.
+             Trace("Loading batch starting with ZINC:$zinc_id.") if T(3);
+             my @attributeData = $fig->query_attributes('$object > ? AND $key = ? ORDER BY $object LIMIT 10000',
+                                                        ["ZINC:$zinc_id", "zinc_name"]);
+             Trace(scalar(@attributeData) . " attribute rows returned.") if T(3);
+             if (! @attributeData) {
+                 # Here there are no attributes left, so we quit the loop.
+                 $done = 1;
+             } else {
+                 # Process the attribute data we've received.
+                 for my $zinc_data (@attributeData) {
+                     # The ZINC ID is found in the first return column, prefixed with the word ZINC.
+                     if ($zinc_data->[0] =~ /^ZINC:(\d+)$/) {
+                         $zinc_id = $1;
+                         # Check for a duplicate.
+                         if ($zinc_id eq $last_zinc_id) {
+                             $loadLigand->Add("duplicate");
+                         } else {
+                             # Here it's safe to output the ligand. The ligand name is the attribute value
+                             # (third column in the row).
+                             $loadLigand->Put($zinc_id, $zinc_data->[2]);
+                             # Insure we don't try to add this ID again.
+                             $last_zinc_id = $zinc_id;
+                         }
+                     } else {
+                         Trace("Invalid zinc ID \"$zinc_data->[0]\" in attribute table.") if T(0);
+                         $loadLigand->Add("errors");
+                     }
+                 }
+             }
+         }
+         Trace("Ligands loaded.") if T(2);
+     }
+     # Finish the load.
+     my $retVal = $self->_FinishAll();
+     return $retVal;
+ }
  =head2 Internal Utility Methods
+ =head3 SpecialAttribute
+ C<< my $count = SproutLoad::SpecialAttribute($id, \@attributes, $idxMatch, \@idxValues, $pattern, $loader); >>
+ Look for special attributes of a given type. A special attribute is found by comparing one of
+ the columns of the incoming attribute list to a search pattern. If a match is found, then
+ a set of columns is put into an output table connected to the specified ID.
+ For example, when processing features, the attribute list we look at has three columns: attribute
+ name, attribute value, and attribute value HTML. The IEDB attribute exists if the attribute name
+ begins with C<iedb_>. The call signature is therefore
+     my $found = SpecialAttribute($fid, \@attributeList, 0, [0,2], '^iedb_', $loadFeatureIEDB);
+ The pattern is matched against column 0, and if we have a match, then column 2's value is put
+ to the output along with the specified feature ID.
+ =over 4
+ =item id
+ ID of the object whose special attributes are being loaded. This forms the first column of the
+ output.
+ =item attributes
+ Reference to a list of tuples.
+ =item idxMatch
+ Index in each tuple of the column to be matched against the pattern. If the match is
+ successful, an output record will be generated.
+ =item idxValues
+ Reference to a list containing the indexes in each tuple of the columns to be put as
+ the second column of the output.
+ =item pattern
+ Pattern to be matched against the specified column. The match will be case-insensitive.
+ =item loader
+ An object to which each output record will be put. Usually this is an B<ERDBLoad> object,
+ but technically it could be anything with a C<Put> method.
+ =item RETURN
+ Returns a count of the matches found.
+ =item
+ =back
+ =cut
+ sub SpecialAttribute {
+     # Get the parameters.
+     my ($id, $attributes, $idxMatch, $idxValues, $pattern, $loader) = @_;
+     # Declare the return variable.
+     my $retVal = 0;
+     # Loop through the attribute rows.
+     for my $row (@{$attributes}) {
+         # Check for a match.
+         if ($row->[$idxMatch] =~ m/$pattern/i) {
+             # We have a match, so output a row. This is a bit tricky, since we may
+             # be putting out multiple columns of data from the input.
+             my $value = join(" ", map { $row->[$_] } @{$idxValues});
+             $loader->Put($id, $value);
+             $retVal++;
+         }
+     }
+     Trace("$retVal special attributes found for $id and loader " . $loader->RelName() . ".") if T(4) && $retVal;
+     # Return the number of matches.
+     return $retVal;
+ }
  =head3 TableLoader
  Create an ERDBLoad object for the specified table. The object is also added to
-Line 1404
+Line 1890
  Name of the table (relation) being loaded.
- =item rowCount (optional)
+ =item ignore
- Estimated maximum number of rows in the table.
+ TRUE if the table should be ignored entirely, else FALSE.
  =item RETURN
-Line 1418
+Line 1904
  sub _TableLoader {
      # Get the parameters.
-     my ($self, $tableName, $rowCount) = @_;
+     my ($self, $tableName, $ignore) = @_;
      # Create the load object.
-     my $retVal = ERDBLoad->new($self->{erdb}, $tableName, $self->{loadDirectory}, $rowCount);
+     my $retVal = ERDBLoad->new($self->{erdb}, $tableName, $self->{loadDirectory}, $self->LoadOnly,
+                                $ignore);
      # Cache it in the loader list.
      push @{$self->{loaders}}, $retVal;
      # Return it to the caller.
-Line 1454
+Line 1941
      my $retVal = Stats->new();
      # Get the loader list.
      my $loadList = $self->{loaders};
+     # Create a hash to hold the statistics objects, keyed on relation name.
+     my %loaderHash = ();
      # Loop through the list, finishing the loads. Note that if the finish fails, we die
-     # ignominiously. At some future point, we want to make the loads restartable.
+     # ignominiously. At some future point, we want to make the loads more restartable.
      while (my $loader = pop @{$loadList}) {
+         # Get the relation name.
+         my $relName = $loader->RelName;
+         # Check the ignore flag.
+         if ($loader->Ignore) {
+             Trace("Relation $relName not loaded.") if T(2);
+         } else {
+             # Here we really need to finish.
+             Trace("Finishing $relName.") if T(2);
          my $stats = $loader->Finish();
+             $loaderHash{$relName} = $stats;
+         }
+     }
+     # Now we loop through again, actually loading the tables. We want to finish before
+     # loading so that if something goes wrong at this point, all the load files are usable
+     # and we don't have to redo all that work.
+     for my $relName (sort keys %loaderHash) {
+         # Get the statistics for this relation.
+         my $stats = $loaderHash{$relName};
+         # Check for a database load.
+         if ($self->{options}->{dbLoad}) {
+             # Here we want to use the load file just created to load the database.
+             Trace("Loading relation $relName.") if T(2);
+             my $newStats = $self->{sprout}->LoadUpdate(1, [$relName]);
+             # Accumulate the statistics from the DB load.
+             $stats->Accumulate($newStats);
+         }
          $retVal->Accumulate($stats);
-         my $relName = $loader->RelName;
          Trace("Statistics for $relName:\n" . $stats->Show()) if T(2);
      }
      # Return the load statistics.
      return $retVal;
  }
+ =head3 GetGenomeAttributes
+ C<< my $aHashRef = GetGenomeAttributes($fig, $genomeID, \@fids); >>
+ Return a hash of attributes keyed on feature ID. This method gets all the NMPDR-related
+ attributes for all the features of a genome in a single call, then organizes them into
+ a hash.
+ =over 4
+ =item fig
+ FIG-like object for accessing attributes.
+ =item genomeID
+ ID of the genome who's attributes are desired.
+ =item fids
+ Reference to a list of the feature IDs whose attributes are to be kept.
+ =item RETURN
+ Returns a reference to a hash. The key of the hash is the feature ID. The value is the
+ reference to a list of the feature's attribute tuples. Each tuple contains the feature ID,
+ the attribute key, and one or more attribute values.
+ =back
+ =cut
+ sub GetGenomeAttributes {
+     # Get the parameters.
+     my ($fig, $genomeID, $fids) = @_;
+     # Declare the return variable.
+     my $retVal = {};
+     # Get a list of the attributes we care about.
+     my @propKeys = $fig->get_group_keys("NMPDR");
+     # Get the attributes.
+     my @aList = $fig->get_attributes("fig|$genomeID%", \@propKeys);
+     # Initialize the hash. This not only enables us to easily determine which FIDs to
+     # keep, it insures that the caller sees a list reference for every known fid,
+     # simplifying the logic.
+     for my $fid (@{$fids}) {
+         $retVal->{$fid} = [];
+     }
+     # Populate the hash.
+     for my $aListEntry (@aList) {
+         my $fid = $aListEntry->[0];
+         if (exists $retVal->{$fid}) {
+             push @{$retVal->{$fid}}, $aListEntry;
+         }
+     }
+     # Return the result.
+     return $retVal;
+ }
 ;

 Legend:



Removed from v.1.18
 


changed lines


 
Added in v.1.84
 Legend:



Removed from v.1.18
 


changed lines


 
Added in v.1.84
-Removed from v.1.18
+Added in v.1.84

MCS Webmaster	ViewVC Help
Powered by ViewVC 1.0.3