wrnap 0.11.0 → 0.12.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 29f26144274a4bd80974e243be7cf969b5f2f49a
-  data.tar.gz: 79793475465dacb80644623212f352010a07c910
+  metadata.gz: 5565b1bf13a5afcdda34f97af733e20c56aa8f72
+  data.tar.gz: 69bb1996d55296c10808270587fd8e7b0ba5cc71
 SHA512:
-  metadata.gz: e614c3d69a0dbe91fe8c623c0246e90ddf0fa22fc2f31dccd847fd4bc6394899f6726e606e773927222e6d7707d1261fb2d1f71d952468fe27230cb140704806
-  data.tar.gz: 39640135b688f94e711bc956ea81dad208008a1a0a5bc5ca21ce6515c9641b6d66d1b5a86648fe34a97a8d436385c908b9c75dd2d6aa7a5b6ed71e962eaad545
+  metadata.gz: a09b6dcf15dcac6f804776c136792f93316e294ca1d8c1230140219e5ce44ed3cd3a72c09b192ab34b2d6bb754841bc7b421fd737f05c11ee88aa5640b2eb1d2
+  data.tar.gz: 36a11bd543ade0dc5e0d267e16117a44b1dc5d8afa4d89e8c1921382afdfb0311094a5fa9f7e893397978dcb0c7f12f63a5ab4b447ce3c74e94ffce40f433f38

data/lib/wrnap/etl/infernal.rb

@@ -1,6 +1,10 @@
 module Wrnap
   module Etl
     module Infernal
+      NAME_REGEX   = />>\s+(\S+)(.*\n){3}.*\s(\d+)\s+(\d+)\s+[\+-].*\n/
+      HIT_SEQUENCE = /^.*\d+\s+(.*)\s+\d+\s*$/
+      LOCAL_END    = /\*\[\s*\d+\s*\]\*/
+
       class << self
         def parse_file(file)
           output = File.exist?(file) ? File.read(file) : file
@@ -11,25 +15,28 @@ module Wrnap
             output.
               gsub(/^(.*\n)*Hit alignments:\n/, "").
               gsub(/Internal CM pipeline statistics summary:\n(.*\n)*$/, "").
-              strip.split(?\n).reject(&:empty?).each_slice(10).map { |lines| parse_hit(lines.join(?\n)) }
+              strip.split(?\n).reject(&:empty?).each_slice(10).map { |lines| parse_hit(lines.join(?\n)) }.compact
           end
         end
 
         def parse_hit(output)
-          name = if output =~ (identifier_regex = />>\s+(\S+)(.*\n){3}.*\s(\d+)\s+(\d+)\s+[\+-].*\n/)
+          name = if output =~ NAME_REGEX
             # This is a pretty fancy regex, and there's no guarantee that the data has this info, so let's just test the waters here.
-            _, id, _, seq_from, seq_to, _ = output.match(identifier_regex).to_a
+            _, id, _, seq_from, seq_to, _ = output.match(NAME_REGEX).to_a
             "%s %d %d" % [id.split(?|).last, seq_from, seq_to]
           end
 
-          Stockholm.fit_structure_to_sequence(
-            *output.split(?\n).as do |infernal|
-              [
-                infernal.select { |line| line =~ /^.*\d+\s+.*\s+\d+\s*$/ }.last.match(/^.*\s+(\S+)\s+\d+\s*$/)[1].upcase.gsub(/[^AUGC]/, ?.),
-                convert_infernal_to_dot_bracket(infernal.find { |line| line =~ /CS\s*$/ }.gsub(/\s+CS\s*$/, "").strip)
-              ]
-            end
-          ).tap { |rna| rna.comment = name if name }
+          unless (hit_sequence = pull_infernal_hit_sequence(output)) =~ LOCAL_END
+            Stockholm.fit_structure_to_sequence(hit_sequence, pull_infernal_hit_structure(output)).tap { |rna| rna.comment = name if name }
+          end
+        end
+
+        def pull_infernal_hit_sequence(output)
+          output.split(?\n).select { |line| line =~ HIT_SEQUENCE }.last.match(HIT_SEQUENCE)[1].upcase
+        end
+
+        def pull_infernal_hit_structure(output)
+          convert_infernal_to_dot_bracket(output.split(?\n).find { |line| line =~ /CS\s*$/ }.gsub(/\s+CS\s*$/, "").strip)
         end
 
         def convert_infernal_to_dot_bracket(structure)

data/lib/wrnap/global/rna/extensions.rb

@@ -7,7 +7,6 @@ module Wrnap
           base.extend(ClassMethods)
           base.extend(OneStructureBasedMethods)
           base.extend(TwoStructureBasedMethods)
-
           base.class_eval do
             OneStructureBasedMethods.public_instance_methods.each do |class_method|
               define_method(class_method) do |*args|
@@ -55,18 +54,6 @@ module Wrnap
         end
 
         module OneStructureBasedMethods
-          def helices(structure)
-            array = base_pairs(structure).sort_by(&:first).map(&:to_a)
-
-            unless array.empty?
-              array[1..-1].inject([[array.first]]) do |bins, (i, j)|
-                bins.tap { bins[-1][-1] == [i - 1, j + 1] ? bins[-1] << [i, j] : bins << [[i, j]] }
-              end
-            else
-              []
-            end
-          end
-
           def max_bp_distance(structure)
             base_pairs(structure).count + ((structure.length - 3) / 2.0).floor
           end

data/lib/wrnap/global/rna/helix.rb

@@ -0,0 +1,36 @@
+module Wrnap
+  module Global
+    class Rna
+      module HelixFunctions
+        def helices
+          array = base_pairs.sort_by(&:first).map(&:to_a)
+
+          unless array.empty?
+            array[1..-1].inject([[array.first]]) do |bins, (i, j)|
+              bins.tap { bins[-1][-1] == [i - 1, j + 1] ? bins[-1] << [i, j] : bins << [[i, j]] }
+            end
+          else
+            []
+          end
+        end
+
+        def collapsed_helices
+          helices.map { |((i, j), *rest)| Helix.new(i, j, rest.length + 1) }
+        end
+      end
+
+      class Helix
+        attr_reader :i, :j
+        attr_accessor :length
+
+        def initialize(i, j, length)
+          @i, @j, @length = i, j, length
+        end
+
+        def name
+          "(%d, %d)" % [i, j]
+        end
+      end
+    end
+  end
+end

data/lib/wrnap/global/rna/tree.rb

@@ -0,0 +1,85 @@
+module Wrnap
+  module Global
+    class Rna
+      module TreeFunctions
+        def with_tree
+          meta_rna { |metadata| tree(TreePlanter.new(metadata.__rna__)) }
+        end
+
+        def trunk; md[:tree] || with_tree.trunk; end
+      end
+
+      class TreePlanter
+        attr_reader :rna, :root
+
+        def initialize(rna, tree = false)
+          @rna  = rna
+          @root = tree || build_tree
+        end
+
+        def build_tree
+          rna.collapsed_helices.inject(Tree::TreeNode.new(:root, rna)) do |tree, helix|
+            node = Tree::TreeNode.new(helix.name, helix)
+
+            if tree.is_root?
+              tree << node
+            elsif helix.i > tree.content.j
+              # It's a sibling, pop up until we're at its parent node.
+              tree = tree.parent until tree.is_root? || tree.content.j > helix.i
+              node.tap { tree << node }
+            elsif helix.j < tree.content.j
+              # Going deeper.
+              tree << node
+            end
+          end.root
+        end
+
+        def coalesce
+          self.class.new(rna, root.dup).tap { |tree| tree.merge_interior_loops! }
+        end
+
+        def coalesce!
+          tap { merge_interior_loops! }
+        end
+
+        def merge_interior_loops!
+          root.tap do
+            self.class.postorder_traversal(root) do |node|
+              if node.children.count == 1 && !node.is_root?
+                child = node.children.first
+                node.parent.add(child)
+                node.remove_from_parent!
+              end
+            end
+          end
+        end
+
+        def depth_signature
+          root.map(&:node_depth)
+        end
+
+        def pp
+          root.print_tree and nil
+        end
+
+        def inspect
+          "#<TreePlanter: %s>" % depth_signature.inspect
+        end
+
+        alias :to_s :inspect
+
+        class << self
+          def preorder_traversal(node, &block)
+            node.children.map { |child| preorder_traversal(child, &block) }
+            yield node
+          end
+
+          def postorder_traversal(node, &block)
+            node.children.map { |child| postorder_traversal(child, &block) }
+            yield node
+          end
+        end
+      end
+    end
+  end
+end

data/lib/wrnap/global/rna.rb

@@ -1,9 +1,11 @@
 module Wrnap
   module Global
     class Rna
-      extend Forwardable
+      extend  Forwardable
       include Extensions
       include Metadata
+      include TreeFunctions
+      include HelixFunctions
 
       CANONICAL_BASES = Set.new << Set.new([?G, ?C]) << Set.new([?A, ?U]) << Set.new([?G, ?U])
 
@@ -175,7 +177,7 @@ module Wrnap
           ("#{seq[0, 20]   + (seq.length > 20   ? '... [%d]' % seq.length : '')}" if seq   && !seq.empty?),
           ("#{str_1[0, 20] + (str_1.length > 20 ? ' [%d]'    % seq.length : '')}" if str_1 && !str_1.empty?),
           ("#{str_2[0, 20] + (str_2.length > 20 ? ' [%d]'    % seq.length : '')}" if str_2 && !str_1.empty?),
-          (md.to_json unless md.empty?),
+          (md.inspect unless md.empty?),
           (name ? name : "#{self.class.name}")
         ].compact.join(", ")
       end

data/lib/wrnap/version.rb

@@ -1,3 +1,3 @@
 module Wrnap
-  VERSION = "0.11.0"
+  VERSION = "0.12.0"
 end

data/lib/wrnap.rb

@@ -1,6 +1,7 @@
 require "yaml"
 require "benchmark"
 require "set"
+require "tree"
 require "shuffle"
 require "rinruby"
 require "tempfile"
@@ -15,6 +16,8 @@ require "active_support/core_ext/class"
 require "wrnap/version"
 require "wrnap/global/rna/extensions"
 require "wrnap/global/rna/metadata"
+require "wrnap/global/rna/tree"
+require "wrnap/global/rna/helix"
 require "wrnap/global/rna"
 require "wrnap/global/rna/context"
 require "wrnap/global/chainer"

data/wrnap.gemspec

@@ -30,4 +30,5 @@ Gem::Specification.new do |spec|
   spec.add_runtime_dependency "bio",           "~> 1.4"
   spec.add_runtime_dependency "bio-stockholm", "~> 0.0.1"
   spec.add_runtime_dependency "entrez",        "~> 0.5"
+  spec.add_runtime_dependency "rubytree",      "~> 0.9"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: wrnap
 version: !ruby/object:Gem::Version
-  version: 0.11.0
+  version: 0.12.0
 platform: ruby
 authors:
 - Evan Senter
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-07-16 00:00:00.000000000 Z
+date: 2014-07-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -136,6 +136,20 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '0.5'
+- !ruby/object:Gem::Dependency
+  name: rubytree
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '0.9'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '0.9'
 description: ''
 email:
 - evansenter@gmail.com
@@ -157,7 +171,9 @@ files:
 - lib/wrnap/global/rna.rb
 - lib/wrnap/global/rna/context.rb
 - lib/wrnap/global/rna/extensions.rb
+- lib/wrnap/global/rna/helix.rb
 - lib/wrnap/global/rna/metadata.rb
+- lib/wrnap/global/rna/tree.rb
 - lib/wrnap/global/runner.rb
 - lib/wrnap/graphing/r.rb
 - lib/wrnap/package/base.rb