rbbt-dm 1.1.8 → 1.1.9

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 2d622e348e2b667b9a7a50d72b838b052a901ee1
-  data.tar.gz: 52aa969ad1244db73caa9869c7aac95ba676d8ea
+  metadata.gz: f4d951189411296e243b213618e445291596cb53
+  data.tar.gz: d7afae72bcaa8b8ee82aaac16d6e83d742386772
 SHA512:
-  metadata.gz: ffc9db25cf186b73e846075f71704c48e2cd0287543d85c949a7ebaeef272351bdd064ed110dcdab6acbccf1e2819133e7a0a7ee74c3a65e7cfc2a42d9c3e070
-  data.tar.gz: a17ba0299410b15b3ca4baedae8b2b85821782f3217809a99f7353f09249a90b86bd4005b30413d8612700cb05d15f4d38f5521976d7de2704bc52c2c2a1fca8
+  metadata.gz: 0354f0e2134e14e836e93648a7ac64a32f79e652198bbaa50458cd780ca4cea3b256d2fd57d486d6a5fc5c90299097afb245f6eb289c1cf3f104b62a57f43f1f
+  data.tar.gz: d356f93d961b8681b03ea9af236d8c6c2a57da222004ec05e8647d49f7e548f26ea1f895446f997df79b9306a986b82a50fabefd4e3f0df4aef8cc837c2fd42a

data/lib/rbbt/matrix/barcode.rb

@@ -9,6 +9,19 @@ source('#{Rbbt.share.R['barcode.R'].find}')
 rbbt.GE.barcode(#{ R.ruby2R self.data_file }, #{ R.ruby2R outfile }, #{ R.ruby2R factor })
     EOF
 
-    R.run(cmd, :stderr => true)
+    R.run(cmd)
   end
+
+  def activity_cluster(outfile, factor = 2)
+
+    FileUtils.mkdir_p File.dirname(outfile) unless outfile.nil? or File.exists? File.dirname(outfile)
+    cmd =<<-EOF
+source('#{Rbbt.share.R['barcode.R'].find}')
+rbbt.GE.activity_cluster(#{ R.ruby2R self.data_file }, #{ R.ruby2R outfile }, #{R.ruby2R value_type})
+    EOF
+
+    R.run(cmd)
+  end
+
+
 end

data/lib/rbbt/matrix/differential.rb

@@ -1,3 +1,5 @@
+require 'rbbt/util/R'
+
 class Matrix
   def differential(main, contrast, path = nil)
     if Array === main and Array === contrast
@@ -7,17 +9,31 @@ class Matrix
     end
 
     name = data_file =~ /:>/ ? File.basename(data_file) : data_file
-    Persist.persist(name, :tsv, :update => true, 
+    Persist.persist(name, :tsv, :persist => :update,
                     :other => {:main => main_samples, :contrast => contrast_samples}, 
                     :prefix => "Diff", :dir => Matrix.matrix_dir.differential, :no_load => true) do |file|
 
-      log2 = value_type.nil? or value_type == "count"
-      log2 = false
-      two_channel = false
-      FileUtils.mkdir_p File.dirname(file) unless file.nil? or File.exists? File.dirname(file)
-      cmd = "source('#{Rbbt.share.R["MA.R"].find}'); rbbt.dm.matrix.differential(#{ R.ruby2R data_file }, main = #{R.ruby2R(main_samples)}, contrast = #{R.ruby2R(contrast_samples)}, log2=#{ R.ruby2R log2 }, outfile = #{R.ruby2R path}, key.field = #{R.ruby2R format}, two.channel = #{R.ruby2R two_channel})"
-      R.run(cmd)
-      nil
+      raise if file.nil?
+
+        log2 = value_type.nil? or value_type == "count"
+        log2 = false
+        two_channel = false
+        FileUtils.mkdir_p File.dirname(file) unless file.nil? or File.exists? File.dirname(file)
+
+        cmd = <<-EOS
+
+source('#{Rbbt.share.R["MA.R"].find}')
+
+data = rbbt.dm.matrix.differential(#{ R.ruby2R data_file }, 
+  main = #{R.ruby2R(main_samples)}, 
+  contrast = #{R.ruby2R(contrast_samples)}, 
+  log2=#{ R.ruby2R log2 }, 
+  outfile = #{R.ruby2R file}, 
+  key.field = #{R.ruby2R format}, 
+  two.channel = #{R.ruby2R two_channel})
+        EOS
+
+        R.run(cmd, :monitor => true)
     end
   end
 end

data/lib/rbbt/matrix/knowledge_base.rb

@@ -0,0 +1,92 @@
+require 'rbbt/matrix'
+require 'rbbt/knowledge_base'
+require 'rbbt/matrix/barcode'
+
+class KnowledgeBase
+  attr_accessor :matrix_registry
+  def matrix_registry=(new)
+    @matrix_registry = IndiferentHash.setup(new)
+  end
+
+  def matrix(name)
+    matrix, options = @matrix_registry[name]
+
+    return matrix if Matrix === matrix
+
+    Path.setup(matrix) if not Path === matrix and File.exists? matrix
+
+    raise "Registered matrix is strange: #{Misc.fingerprint matrix}" unless Path === matrix
+
+    path = matrix
+
+    raise "Registered path not found: #{path.find}" unless path.exists?
+    
+    if path.find.directory?
+      data, labels, value_type, format, organism, identifiers = Misc.process_options options, :data, :labels, :value_type, :format, :organism, :identifiers 
+
+      data ||= path.data if path.data.exists?
+      data ||= path.values if path.values.exists?
+
+      labels ||= path.labels if path.labels.exists?
+      labels ||= path.samples if path.samples.exists?
+
+      identifiers ||= path.identifiers if path.identifiers.exists?
+
+      value_type = TSV.parse_header(data.find).key_field if data
+      value_type ||= "Unknown ID"
+
+      Matrix.new data, labels, value_type, format, organism, identifiers
+    else
+    end
+  end
+
+  def register_matrix(name, matrix, options = {})
+    options = Misc.add_defaults options, :sample_format => "Sample"
+    sample_format = Misc.process_options options, :sample_format
+
+    @matrix_registry ||= IndiferentHash.setup({})
+    @matrix_registry[name] = [matrix, options]
+
+
+    register name do
+      matrix = matrix(name)
+      TSV.read_matrix matrix.data_file, sample_format
+    end
+
+    register name.to_s + '_activity' do
+      matrix = matrix(name)
+      TmpFile.with_file do |tmpfile|
+        matrix.activity_cluster(tmpfile)
+        tsv = TSV.open(TSV.read_matrix(tmpfile, sample_format))
+        tsv.identifiers ||= matrix.data_file.identifier_files.first
+        tsv.identifiers = tsv.identifiers.find if tsv.identifiers.respond_to? :find
+        
+        tsv = tsv.add_field "Activity" do |k,p|
+          samples, values = p
+          values = values.collect{|v| v.to_i }
+          new_values = case Misc.max(values) 
+                       when 1
+                         [''] * samples.length
+                       when 2
+                         values.collect{|v| v == 2 ? "active" : '' }
+                       else
+                         values.collect{|v|
+                           case v
+                           when 1
+                             "inactive"
+                           when 2
+                             ''
+                           else
+                             "active"
+                           end
+                         }
+          end
+        end
+
+        tsv
+      end
+    end
+  end
+
+end
+  

data/lib/rbbt/matrix.rb

@@ -10,17 +10,23 @@ class Matrix
     end
   end
 
-  attr_accessor :data_file, :labels, :value_type, :format, :organism
+  attr_accessor :data_file, :labels, :value_type, :format, :organism, :identifiers
   def initialize(data_file, labels, value_type, format, organism=nil, identifiers=nil)
     @data_file = data_file
     @labels = labels
     @value_type = value_type
     @format = format
+    @format ||=  begin
+                   _header ||= TSV.parse_header(@data_file)
+                   _header.key_field || "ID"
+                 end
     @organism = organism 
+    _header = nil
     @organism ||=  begin
-                     TSV.parse_header(@data_file).namespace || "Hsa"
+                     _header ||= TSV.parse_header(@data_file)
+                     _header.namespace || "Hsa"
                    end
-    @identifiers = identifiers || Organism.identifiers(organism)
+    @identifiers = identifiers 
   end
 
   def samples
@@ -31,6 +37,17 @@ class Matrix
     @subsets ||= begin
                    subsets = {}
                    case @labels
+                   when Path
+                     labels = @labels.tsv
+                     factors = labels.fields
+                     labels.through do |sample,values|
+                       factors.zip(values).each do |factor,value|
+                         subsets[factor] ||= {}
+                         subsets[factor][value] ||= []
+                         subsets[factor][value] << sample
+                       end
+                     end
+
                    when TSV
                      factors = @labels.fields
                      @labels.through do |sample,values|
@@ -42,7 +59,7 @@ class Matrix
                      end
                    when Hash
                      @labels.each do |factor,info|
-                       subsets[factors] ||= []
+                       subsets[factors] ||= {}
                        info.each do |value, samples|
                          subsets[factors][value] = case samples
                                                    when Array 
@@ -56,16 +73,29 @@ class Matrix
                        end
                      end
                    end
-                   subsets
+
+                   clean_subsets = {}
+                   subsets.each do |factor,values|
+                     next if values.nil? or values.size < 2
+                     values.each do |level,samples|
+                       next if samples.nil? or samples.length < 2
+                       clean_subsets[factor] ||= {}
+                       clean_subsets[factor][level] = samples
+                     end
+                   end
+
+                   clean_subsets
                  end
   end
 
   def comparison(main, contrast, subsets = nil)
+    subsets ||= self.subsets
 
     if main.index "="
       main_factor, main_value = main.split "=" 
       raise ParameterException, "Main selection not understood" if subsets[main_factor].nil? or subsets[main_factor][main_value].nil?
-      main_samples = subsets[main_factor][main_value].split ','
+      value = subsets[main_factor][main_value]
+      main_samples = String === value ? value.split(',') : value
     else
       main_samples = main.split(/[|,\n]/)
     end
@@ -74,17 +104,20 @@ class Matrix
       if contrast.index "="
         contrast_factor, contrast_value = contrast.split "=" 
         raise ParameterException, "Contrast selection not understood" if subsets[contrast_factor].nil? or subsets[contrast_factor][contrast_value].nil?
-        contrast_samples = subsets[contrast_factor][contrast_value].split ','
+        value = subsets[contrast_factor][contrast_value]
+        contrast_samples = String === value ? value.split(',') : value
       else
         contrast_samples = contrast.split(/[|,\n]/)
       end
     else
       if subsets and defined? main_factor
-        contrast_samples = subsets[main_factor].values.collect{|s| s.split ',' }.flatten.uniq - main_samples
+        contrast_samples = subsets[main_factor].values.flatten.collect{|s| s.split ',' }.flatten.uniq - main_samples
       else
         contrast_samples = samples - main_samples
       end
     end
+    main_samples = main_samples.compact.reject{|m| m.empty? }.collect{|m| m.strip }
+    contrast_samples = contrast_samples.compact.reject{|m| m.empty? }.collect{|m| m.strip }
 
     [main_samples, contrast_samples]
   end
@@ -96,9 +129,12 @@ class Matrix
     name = data_file =~ /:>/ ? File.basename(data_file) : data_file
 
     file = Persist.persist(data_file, :tsv, :prefix => "Gene", :dir => Matrix.matrix_dir.values, :no_load => true) do
-      identifiers = [Organism.identifiers(organism), @identifiers, identifiers].compact.uniq
 
-      data_file.tsv(:cast => :to_f).change_key("Ensembl Gene ID", :identifiers => identifiers) do |v|
+      data = data_file.tsv(:cast => :to_f)
+
+      identifiers = [identifiers, @identifiers, data.identifiers, Organism.identifiers(organism)].flatten.compact.uniq
+
+      data.change_key("Ensembl Gene ID", :identifiers => identifiers) do |v|
         Misc.mean(v.compact)
       end
     end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rbbt-dm
 version: !ruby/object:Gem::Version
-  version: 1.1.8
+  version: 1.1.9
 platform: ruby
 authors:
 - Miguel Vazquez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-09-16 00:00:00.000000000 Z
+date: 2014-09-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rbbt-util
@@ -94,6 +94,7 @@ files:
 - lib/rbbt/matrix.rb
 - lib/rbbt/matrix/barcode.rb
 - lib/rbbt/matrix/differential.rb
+- lib/rbbt/matrix/knowledge_base.rb
 - lib/rbbt/network/paths.rb
 - lib/rbbt/plots/bar.rb
 - lib/rbbt/plots/heatmap.rb