new scaffold for a Gadfly adaptor layer

lib/Bio/DB/Gadfly/Adaptor.pm

@@ -113,6 +113,29 @@ sub segment {
   return;
 }
 
+# things that user could ask for (ss = "search string")
+# WARNING WARNING WARNING: gadfly is interbase coordinates
+# 1. gene
+#      name  ("CG" name)
+#      symbol (fruitfly nomenclature official name)
+#      synonym (may return multiple results)
+# SOLUTION: fetch object via lget_Gene with constraint name|symbol|synonym=>ss
+#           get start, end, and seq_object and throw out rest
+# 2. primary sequence
+#      BUT, what user really wants is alignment of this primary sequence
+#            to genome (via a resultset)
+# SOLUTION: fetch object via lget_ResultSet with constraint subject_seq=>ss
+#           get start, end and seq_object and throw out rest
+# 3. chromosome arm
+# SOLUTION: fetch primary seq from database
+#           get start, end and seq_object and throw out rest
+#
+# 4. cytological range
+# just feed into CytoRange object
+# SOLUTION: if range given construct a CytoRange object (BioModel::CytoRange->new(-range_string=>$string)
+#           call start_band() and end_band() to give band objects.
+#           get_CytoBand({name=>$range->start_band()->name}) this will give coordinates
+
 =head2 @classes = $db->classes;
 
 Return the namespaces known to the database.
@@ -121,9 +144,20 @@ Return the namespaces known to the database.
 
 # hard-coded classes
 sub classes {
-  return qw(name symbol);
+  return qw(name symbol synonym sequence arm);
 }
 
+# this is not very useful and shouldn't be using SQL
+# useful:
+
+# object      constraint
+# ------      ----------
+# gene        comment
+#             transcript_comment
+#             term
+#             dbxref
+# primary_seq description
+
 =head2 @result = $db->search_notes('search string')
 
 Perform a text search on database.  Result is an array of [$name,$description,$score].

lib/Bio/DB/Gadfly/Segment.pm

@@ -30,6 +30,9 @@ BioModel::AnnotatedSeq.  We evaluate in a lazy way.
 
 =cut
 
+# INSTEAD OF LAZY EVALUATION, GIVE AS OPTION ALTERNATIVE IN WHICH CALL GET_ANNOTATEDSEQ() in new()
+# store result and then grep for features of certain types
+
 sub new {
   my $class = shift;
   my ($name,$start,$end,$gx,$other) = rearrange([qw(NAME START END GX)],@_);
@@ -67,13 +70,26 @@ Return the end coordinate using bioperl conventions
 
 sub end { shift->{end} }
 
-=head2 @features = $db->features(@args);
+=head2 @features = $db->overlapping_features(@args);
 
 Return a list of Bio::SeqFeatureI objects that overlap this segment.
 
 =cut
 
-sub features {
+# gadfly types
+# annotation type space
+#     biological names like "gene", "snRNA", etc
+#     annotation:gene, annotation:snRNA
+# analysis type space
+#     "blastx against nonfly"...
+#     analysis:blastx_against_nonfly
+# scaffold type space
+#     scaffold(:nothing)
+
+# chris will implement a to_bioperl_feature() method in the
+# genes, analyses and scaffolds
+
+sub overlapping_features {
   my $self   = shift;
   my ($types,$rangetype,$attributes,$iterator);
 
@@ -95,7 +111,7 @@ sub features {
   # everything else is handled as an analysis
   my %types          = map {$_=>1} @$types;
   my @gene_types     = grep {/^transcript(:|$ )/x} keys %types;
-  my @segment_types  = grep {/^segment(:|$ )/x} keys %types;
+  my @scaffold_types = grep {/^scaffold(:|$ )/x} keys %types;
   my %special_types  = map {$_=>1} (@gene_types,@segment_types);
   my @analysis_types = grep {!$special_types{$_}} keys %types;
 
@@ -104,30 +120,38 @@ sub features {
   my @features;
 
   if ($all_types or @gene_types) {
-    push @features,$self->_get_genes($segment,\@gene_types);
+    push @features,map {$_->to_bioperl_feature} $self->_get_genes($segment,\@gene_types);
   }
-  if ($all_types or @segment_types) {
-    push @features,$self->_get_segments($segment,\@segment_types);
+  if ($all_types or @scaffold_types) {
+    push @features,map {$_->to_bioperl_feature} $self->_get_scaffolds($segment,\@scaffold_types);
   }
   if ($all_types or @analysis_types) {
-    push @features,$self->_get_analyses($segment,\@analysis_types);
+    push @features,map {$_->to_bioperl_feature} $self->_get_analyses($segment,\@analysis_types);
   }
 
   # iteration doesn't really help us with gadfly, but we support it because
   # it's part of the interface.
   if ($iterator) {
-    return Bio::DB::Gadfly::SegmentIterator->new(\@analysis_types);
+    return Bio::DB::Gadfly::SegmentIterator->new(\@features);
   }
+
   elsif ($callback) {
-    my $continue = 1;
-    map { $continue &&= $callback->($_)} @features;
+    foreach (@features) {
+      my $result = $callback->($_);
+      last unless $result;
+    }
     return scalar @features;
   }
+
   else {
     return @features;
   }
 }
 
+#NOTE: the Bio::SeqIO::game class should be able to dump out GAME/XML
+# from these objects by using the gx() method and the start/end position
+# information.
+
 =head2 $gx_adaptor = $db->gx([$new_adaptor])
 
 Get or set the underlying GxAdapter (note spelling difference).
@@ -144,12 +168,25 @@ sub gx {
 sub _get_genes {
   my $self = shift;
   my ($segments,$types) = @_;
+  # $gx->lget_Genes() for lazy evaluation
+  # $as->gene_list() for non-lazy evaluation
 }
 
-sub _get_segments {
+sub _get_scaffolds {
+  # lazy:
+  # @scaffolds = $gx->lget_ResultSet({constraints...,analysis_program=>'tiling_path'});
+  # non-lazy:
+  # $analysis = $as->get_analysis('tiling_path')
+  # @scaffolds = $analysis->result_set_list()
+  # @scaffolds are result sets
 }
 
 sub _get_analyses {
+  # lazy:
+  # @analyses = $gx->lget_ResultSet({constraints...,analysis_types=>['types'...]})
+  # non-lazy
+  # $all_analyses = $as->analysis_list()
+  # foreach ($all_analyses) { next unless $_->get_property('type'); do something }
 }
 
 sub _tosegment {