matr 0.0.2 → 0.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 0aefd54bc8a7c57da41e3843bbc8cd9e29885457
-  data.tar.gz: 3f6a1f167731135a93423bcb076a99259a22a5de
+  metadata.gz: 27143ae159138fad54cd2d4c85138c47dbacf90a
+  data.tar.gz: d35cebbbe078e2a3689c9f32ba1c093c1befa2ca
 SHA512:
-  metadata.gz: 47c0ca490761b79d628b72ba7ae805450844da7ec550284207fe4cea78e497cf239f71faf9ee5de070b319e9137bd869997c3776bd3d764db0ecf5a01f722326
-  data.tar.gz: 0baeb31d028459fba8ab48a409e3f5666c1c3d4449504b92e32d233b5291acf3288801404d628d36acedccc4657cdebb1a3cbf4000a608bec0d74d7bcfa074ea
+  metadata.gz: '08c61835a0633930babf25435a2ee4d96a5f2444003ff1ce1ced4ce426d98eb18459738619007d9d2c5b84cdb0a2586d35ed297061aeb6be8cff6978df57526e'
+  data.tar.gz: fbee20edd8a637b072b3f6e5a34964f11131744722e316ad9e59567c276bcd7143a960a50ad6953e123aea855ae763854370f663368ee7f5e3a0afc1a8bd0c32

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    matr (0.0.1)
+    matr (0.1.0)
       trollop (~> 2.1, >= 2.1.3)
 
 GEM

data/exe/matr

@@ -2,6 +2,8 @@
 
 Signal.trap("PIPE", "EXIT")
 
+require "matr"
+
 require "set"
 require "trollop"
 
@@ -10,89 +12,33 @@ opts = Trollop.options do
 
   Take a three-column matrix file and ensure that it is symmetric for input into R.
 
-  Options:
-  EOS
+  *Note* -- I ALWAYS enforce symmetry of the output.
 
-  opt(:infile, "Input file", type: :string)
-
-  opt(:self_score, "Score for self connections", default: 100)
-  opt(:missing_score, "Score for missing connections", default: 0)
-  opt(:output_style, "Output style [wide,long]", default: "wide")
-end
-
-infile = opts[:infile]
-self_score = opts[:self_score]
-missing_score = opts[:missing_score]
-output_style = opts[:output_style]
-
-unless File.exist? infile
-  abort "FATAL -- infile '#{infile}' does not exist"
-end
-
-unless output_style == "wide" || output_style == "long"
-  abort "FATAL -- output style must be either wide or long"
-end
+  Option details
+  --------------
 
-graph = Hash.new { |ht, k| ht[k] = Hash.new }
-all_keys = Set.new
-header   = nil
+  --self-scores: I will add a self score of this value if one is not already specified.
+  --ensure-self-scores: I will ensure all self scores equal the value set in --self-scores even if a self-score is present but set to something else.
 
-File.open(infile, "rt").each_line.with_index do |line, idx|
-  line.chomp!
+  Things I do
+  ------------------
 
-  if idx.zero?
-    header = line
-  else
-    source, target, score = line.split "\t"
+    - Raise error if a source-target pair is specified twice with a different score
+    - Raise error if source-target and target-source are both specified and they don't match.
 
-    all_keys << source << target
-
-    # If you have duplicates with different scores, that's an error: a
-    # => b = 10 and a => b = 20
-    if graph[source].has_key?(target) && graph[source][target] != score
-      abort "source--target (#{source}--#{target}) was repeated with a different score"
-    end
-
-    graph[source][target] = score
+  Options:
+  EOS
 
-    # Since you might have a files that specifies a => b = 10, and b
-    # => a = 10, that would be okay.  But if you have a => b = 10, and
-    # b => a = 5, then that would be an error.
-    if graph[target].has_key?(source) && graph[target][source] != score
-      abort "target--source (#{target}--#{source}) is specified seperately from source--target (#{source}--#{target}), BUT there scores do not match.  This means your matrix is NOT symmetric."
+  opt(:infile, "Input file", type: :string)
 
-    end
+  opt(:missing_score, "Score for missing connections (unless they are already put in because of symmetry)", default: 0)
 
-    graph[target][source] = score
+  # opt(:ensure_symmetry, "This flag will ensure that your output will be a symmetric matrix.", default: true)
 
-    unless graph[target].has_key? target
-      graph[target][target] = self_score
-    end
+  opt(:self_score, "Score for self connections", default: 100)
+  opt(:ensure_self_scores, "This flag will ensure that self hit scores all match the value of --self-score option", default: true)
 
-    unless graph[source].has_key? source
-      graph[source][source] = self_score
-    end
-  end
+  opt(:output_style, "Output style [wide,long]", default: "wide")
 end
 
-if output_style == "long"
-  puts header
-
-  all_keys.each do |source|
-    all_keys.each do |target|
-      puts [source, target, graph[source][target] || missing_score].join "\t"
-    end
-  end
-else
-  all_keys = all_keys.sort
-
-  puts ["", all_keys].join "\t"
-
-  all_keys.each do |source|
-    row = all_keys.map do |target|
-      graph[source][target] || missing_score
-    end
-
-    puts [source, row].join "\t"
-  end
-end
+Matr.main opts

data/lib/matr.rb

@@ -1,5 +1,96 @@
 require "matr/version"
 
 module Matr
-  # Your code goes here...
+  def self.default_opt opts, key, default
+    opts.has_key?(key) ? opts[key] : default
+  end
+
+  def self.main opts
+
+    infile             = opts[:infile]
+    self_score         = self.default_opt opts, :self_score, 100
+    missing_score      = self.default_opt opts, :missing_score, 0
+    output_style       = self.default_opt opts, :output_style, "wide"
+    ensure_symmetry    = true # self.default_opt opts, :ensure_symmetry, true
+    ensure_self_scores = self.default_opt opts, :ensure_self_scores, true
+
+    unless File.exist? infile
+      raise StandardError, "FATAL -- infile '#{infile}' does not exist"
+    end
+
+    unless output_style == "wide" || output_style == "long"
+      raise StandardError, "FATAL -- output style must be either wide or long"
+    end
+
+    graph    = Hash.new { |ht, k| ht[k] = Hash.new }
+    all_keys = Set.new
+    header   = nil
+
+    File.open(infile, "rt").each_line.with_index do |line, idx|
+      line.chomp!
+
+      if idx.zero?
+        header = line
+      else
+        source, target, score = line.split "\t"
+
+        all_keys << source << target
+
+        # Fix self score if the option is there.
+        if source == target && ensure_self_scores
+          score = self_score
+        end
+
+        # If you have duplicates with different scores, that's an error: a
+        # => b = 10 and a => b = 20
+        if graph[source].has_key?(target) && graph[source][target] != score
+          raise StandardError, "FATAL -- source--target (#{source}--#{target}) was repeated with a different score"
+        end
+
+        graph[source][target] = score
+
+        if ensure_symmetry
+          # Since you might have a files that specifies a => b = 10, and b
+          # => a = 10, that would be okay.  But if you have a => b = 10, and
+          # b => a = 5, then that would be an error.
+          if graph[target].has_key?(source) && graph[target][source] != score
+            raise StandardError, "FATAL -- target--source (#{target}--#{source}) is specified seperately from source--target (#{source}--#{target}), BUT their scores do not match.  This means your matrix is NOT symmetric.  Did you really mean to pass the --ensure-symmetry flag?"
+          end
+
+          graph[target][source] = score
+        end
+
+        # Makes sure these weren't already specified earlier.
+        unless graph[target].has_key? target
+          graph[target][target] = self_score
+        end
+
+        unless graph[source].has_key? source
+          graph[source][source] = self_score
+        end
+      end
+    end
+
+    if output_style == "long"
+      puts header
+
+      all_keys.each do |source|
+        all_keys.each do |target|
+          puts [source, target, graph[source][target] || missing_score].join "\t"
+        end
+      end
+    else
+      all_keys = all_keys.sort
+
+      puts ["", all_keys].join "\t"
+
+      all_keys.each do |source|
+        row = all_keys.map do |target|
+          graph[source][target] || missing_score
+        end
+
+        puts [source, row].join "\t"
+      end
+    end
+  end
 end

data/lib/matr/version.rb

@@ -1,3 +1,3 @@
 module Matr
-  VERSION = "0.0.2"
+  VERSION = "0.1.0"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: matr
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.1.0
 platform: ruby
 authors:
 - Ryan Moore
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2018-08-23 00:00:00.000000000 Z
+date: 2018-08-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: trollop