lederhosen 2.0.6 → 2.0.7

data/lederhosen.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = "lederhosen"
-  s.version = "2.0.6"
+  s.version = "2.0.7"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Austin G. Davis-Richardson"]
-  s.date = "2013-02-27"
+  s.date = "2013-03-11"
   s.description = "Various tools for OTU clustering"
   s.email = "harekrishna@gmail.com"
   s.executables = ["lederhosen"]

data/lib/lederhosen/tasks/otu_filter.rb

@@ -18,57 +18,71 @@ module Lederhosen
 
       ohai "filtering otu file #{input} (reads = #{reads}, samples = #{min_samples})"
 
-      cluster_sample_count = Hash.new { |h, k| h[k] = Hash.new }
+      # make one pass finding which OTUs to keep
+      # create mask that maps which columns correspond to good OTUs
+      # pass over table again printing only those columns
 
-      ohai "loading csv file #{input}"
+      seen = Hash.new { |h, k| h[k] = 0 }
 
-      # slurp up CSV file
-      File.open input do |handle|
+      otu_order = []
+
+      pbar = ProgressBar.new 'counting', File.size(input)
+      total_reads = 0
+
+      File.open(input) do |handle|
         header = handle.gets.strip.split(',')
-        cluster_ids = header[0..-1]
+        header.each { |x| otu_order << x }
+
         handle.each do |line|
+          pbar.set handle.pos
           line = line.strip.split(',')
-          sample_id = line[0].to_sym
-          counts = line[1..-1].map(&:to_i)
-          cluster_ids.zip(counts).each do |cluster, count|
-            cluster_sample_count[cluster][sample_id] = count
+          sample_name = line[0]
+          abunds = line[1..-1].map &:to_i
+          otu_order.zip(abunds) do |o, a|
+            total_reads += a
+            seen[o] += 1 if a >= reads
           end
         end
       end
 
-      ohai "filtering"
+      pbar.finish
 
-      # filter sample_cluster_count
-      # todo: move filtered reads to 'unclassified_reads' classification
-      filtered = cluster_sample_count.reject { |k, v| v.reject { |k, v| v < reads }.size < min_samples }
+      mask = otu_order.map { |x| seen[x] >= min_samples }
 
-      # use functional programming they said
-      # it will make your better they said
-      noise = cluster_sample_count.keys - filtered.keys
+      ohai "found #{otu_order.size} otus, keeping #{mask.count(true)}"
 
-      ohai "saving to #{output}"
+      output = File.open(output, 'w')
 
-      # save the table
-      out = File.open(output, 'w')
-      samples = filtered.values.map(&:keys).flatten.uniq
-      clusters = filtered.keys
-      out.puts "-,#{clusters.join(',')},noise"
+      pbar = ProgressBar.new 'writing', File.size(input)
+      filtered_reads = 0
+      File.open(input) do |handle|
+        header = handle.gets.strip.split(',')
+        header = header.zip(mask).map { |k, m| k if m }.compact
+        output.print header.join(',')
+        output.print ",noise\n" # need a "noise" column
+
+        handle.each do |line|
+          pbar.set handle.pos
+          line = line.strip.split(',')
+
+          sample_name = line[0]
+          counts = line[1..-1].map &:to_i
+
+          kept_counts = counts.zip(mask).map { |c, m| c if m }.compact
+          noise = counts.zip(mask).map { |c, m| c unless m }.compact.inject(:+)
+          filtered_reads += noise
+
+          output.puts "#{sample_name},#{kept_counts.join(',')},#{noise}"
 
-      samples.each do |sample|
-        out.print "#{sample}"
-        clusters.each do |cluster|
-          out.print ",#{filtered[cluster][sample]}"
         end
-        noise_sum = noise.map { |n| cluster_sample_count[n][sample] }.inject(:+)
-        out.print ",#{noise_sum || 0}"
-        out.print "\n"
       end
-      out.close
 
-      ohai "kept #{filtered.keys.size} clusters (#{filtered.keys.size/cluster_sample_count.size.to_f})."
-      kept_reads = filtered.values.map { |x| x.values.inject(:+) }.inject(:+)
-      total_reads = cluster_sample_count.values.map { |x| x.values.inject(:+) }.inject(:+)
-      ohai "kept #{kept_reads}/#{total_reads} reads (#{kept_reads/total_reads.to_f})."
+      pbar.finish
+
+      ohai "kept #{total_reads - filtered_reads}/#{total_reads} reads (#{100*(total_reads - filtered_reads)/total_reads.to_f}%)"
+
+      output.close
+
     end
 
   end

data/lib/lederhosen/version.rb

@@ -3,7 +3,7 @@ module Lederhosen
     MAJOR = 2
     MINOR = 0
     CODENAME = 'Schnittlauchbrot' # changes for minor versions
-    PATCH = 6
+    PATCH = 7
 
     STRING = [MAJOR, MINOR, PATCH].join('.')
   end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: lederhosen
 version: !ruby/object:Gem::Version
-  version: 2.0.6
+  version: 2.0.7
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-02-27 00:00:00.000000000 Z
+date: 2013-03-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: dna
@@ -149,7 +149,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -902984509893215788
+      hash: -4462113867490056177
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements: