iroki 0.0.1

data/lib/iroki/core_ext/file/file.rb

@@ -0,0 +1,64 @@
+# Copyright 2016 Ryan Moore
+# Contact: moorer@udel.edu
+#
+# This file is part of Iroki.
+#
+# Iroki is free software: you can redistribute it and/or modify it
+# under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Iroki is distributed in the hope that it will be useful, but
+# WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Iroki.  If not, see <http://www.gnu.org/licenses/>.
+
+module Iroki
+  module CoreExt
+    module File
+      def check_file arg, which
+        help = " Try iroki --help for help."
+
+        abort_if arg.nil?,
+                 "You must provide a #{which} file.#{help}"
+
+        abort_unless Object::File.exists?(arg),
+                     "The file #{arg} doesn't exist.#{help}"
+
+        arg
+      end
+
+      def parse_name_map fname
+        check_file fname, :name_map
+
+        name_map = {}
+        Object::File.open(fname).each_line do |line|
+          oldname, newname = line.chomp.split "\t"
+
+
+          abort_if oldname.nil? || oldname.empty?,
+                   "Column 1 missing for line: #{line.inspect}"
+
+          abort_if newname.nil? || newname.empty?,
+                   "Column 2 missing for line: #{line.inspect}"
+
+          oldname = clean oldname
+          newname = clean newname
+
+          abort_if name_map.has_key?(oldname),
+                   "#{oldname} is repeated in column 1"
+
+          name_map[oldname] = newname
+        end
+
+        abort_if duplicate_values?(name_map),
+                 "Names in column 2 of name map file must be unique"
+
+        name_map
+      end
+    end
+  end
+end

data/lib/iroki/core_ext/hash/hash.rb

@@ -0,0 +1,28 @@
+# Copyright 2016 Ryan Moore
+# Contact: moorer@udel.edu
+#
+# This file is part of Iroki.
+#
+# Iroki is free software: you can redistribute it and/or modify it
+# under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Iroki is distributed in the hope that it will be useful, but
+# WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Iroki.  If not, see <http://www.gnu.org/licenses/>.
+
+module Iroki
+  module CoreExt
+    module Hash
+      def duplicate_values? hash
+        values = hash.values
+        values.count != 1 && values.count != values.uniq.count
+      end
+    end
+  end
+end

data/lib/iroki/core_ext/string/string.rb

@@ -0,0 +1,34 @@
+module Iroki
+  module CoreExt
+    module String
+
+      def hex? str
+        str.match(/^#[0-9A-Fa-f]{6}$/)
+      end
+
+      def clean str
+        str.gsub(/[^\p{Alnum}_]+/, "_").gsub(/_+/, "_")
+      end
+
+      def has_color? name
+        name.match(/(.*)(\[&!color="#[0-9A-Fa-f]{6}"\])/)
+      end
+      alias already_checked? has_color?
+
+      def clean_name name
+        if name.nil?
+          nil
+        else
+          if (match = has_color? name)
+            name = match[1]
+            color = match[2]
+
+            clean(name) + color
+          else
+            clean(name)
+          end
+        end
+      end
+    end
+  end
+end

data/lib/iroki/main/main.rb

@@ -0,0 +1,232 @@
+# Copyright 2016 Ryan Moore
+# Contact: moorer@udel.edu
+#
+# This file is part of Iroki.
+#
+# Iroki is free software: you can redistribute it and/or modify it
+# under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Iroki is distributed in the hope that it will be useful, but
+# WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Iroki.  If not, see <http://www.gnu.org/licenses/>.
+
+require "bio"
+require "set"
+
+module Iroki
+  module Main
+    def self.main(color_branches: nil,
+                  color_taxa_names: nil,
+                  exact: nil,
+                  remove_bootstraps_below: nil,
+                  color_map_f: nil,
+                  name_map_f: nil,
+                  auto_color: nil,
+                  display_auto_color_options: nil,
+                  newick_f: nil,
+                  out_f: nil)
+
+
+
+      if display_auto_color_options
+        puts "\n  Choices for --auto-color ..."
+        print "  basic, basic_light, basic_dark, funky, funky_light, " +
+              "funky_dark\n\n"
+        exit
+      end
+
+      auto_color_options =
+        ["basic", "basic_light", "basic_dark",
+         "funky", "funky_light", "funky_dark",]
+
+
+      if(!auto_color.nil? &&
+         !auto_color_options.include?(auto_color))
+        puts "\n  Choices for --auto-color ..."
+        print "  basic, basic_light, basic_dark, funky, funky_light, " +
+              "funky_dark\n\n"
+
+        Trollop.die :auto_color, "#{auto_color} is not a valid option"
+      end
+
+      case auto_color
+      when nil
+        auto_colors = BASIC
+      when "basic"
+        auto_colors = BASIC
+      when "basic_light"
+        auto_colors = BASIC_LIGHT
+      when "basic_dark"
+        auto_colors = BASIC_DARK
+      when "funky"
+        auto_colors = FUNKY
+      when "funky_light"
+        auto_colors = FUNKY_LIGHT
+      when "funky_dark"
+        auto_colors = FUNKY_DARK
+      end
+
+      # color_branches = true
+      # color_taxa_names = true
+      # exact = false
+      # color_map_f = "test_files/500.patterns_with_name_map"
+      # name_map_f = "test_files/500.name_map"
+      # ARGV[0] = "test_files/500.zetas.tre"
+      newick = check_file newick_f, :newick
+
+      color_f = nil
+      if color_taxa_names || color_branches
+        color_f = check_file color_map_f, :color_map_f
+      end
+
+      check = color_map_f &&
+              !color_taxa_names &&
+              !color_branches
+
+      abort_if check,
+               "A pattern file was provided without specifying " +
+               "any coloring options"
+
+
+      # if passed color other than one defined, return black
+      black = "#000000"
+      red = "#FF1300"
+      yellow = "#FFD700"
+      blue = "#5311FF"
+      green = "#00FF2C"
+      color2hex = Hash.new "[&!color=\"#{black}\"]"
+      color2hex.merge!({
+                         "black" => "[&!color=\"#{black}\"]",
+                         "red" => "[&!color=\"#{red}\"]",
+                         "blue" => "[&!color=\"#{blue}\"]",
+                         "yellow" => "[&!color=\"#{yellow}\"]",
+                         "green" => "[&!color=\"#{green}\"]"
+                       })
+
+      # check if complementary colors requested
+      if color_f
+        colors = Set.new
+        File.open(color_f).each_line do |line|
+          _, color = line.chomp.split "\t"
+
+          colors << color
+        end
+
+        auto_color = colors.all? { |color| color.match /\A[0-4]\Z/ }
+      end
+
+      # get the color patterns
+      if color_f
+        patterns = {}
+        File.open(color_f).each_line do |line|
+          pattern, color = line.chomp.split "\t"
+
+          color = "black" if color.nil? || color.empty?
+
+          if name_map_f || color_taxa_names || color_branches
+            pattern = clean_name pattern
+          end
+
+          if !exact
+            pattern = Regexp.new pattern
+          end
+
+          if auto_color
+            patterns[pattern] = "[&!color=\"#{auto_colors[color]}\"]"
+          else
+            if hex? color
+              patterns[pattern] = "[&!color=\"#{color}\"]"
+            else
+              patterns[pattern] = color2hex[color]
+            end
+          end
+        end
+      end
+
+      treeio = Bio::FlatFile.open(Bio::Newick, newick)
+
+      newick = treeio.next_entry
+      tree = newick.tree
+
+      # do this first cos everything after will use the "new" names
+      if name_map_f
+        name_map = parse_name_map name_map_f
+
+        tree.collect_node! do |node|
+          unless node.name.nil?
+            # every name is cleaned no matter what
+            node.name = clean node.name
+
+            if name_map.has_key?(node.name)
+              node.name = name_map[node.name]
+            end
+          end
+
+          node
+        end
+      end
+
+      if color_taxa_names
+        leaves = tree.leaves.map do |n|
+          name = clean_name n.name
+
+          if (color = add_color_to_leaf_branch(patterns, name, exact))
+            name + color
+          else
+            name
+          end
+        end
+      else
+        leaves = tree.leaves.map { |n| clean_name n.name }
+      end
+
+      if color_branches
+        total = tree.nodes.count
+        n = 0
+        tree.collect_node! do |node|
+          n += 1
+          $stderr.printf "Node: %d of %d\r", n, total
+
+          color_nodes patterns, tree, node, exact
+        end
+      end
+      $stderr.puts
+
+      if remove_bootstraps_below
+        tree.collect_node! do |node|
+          if node.bootstrap && node.bootstrap < remove_bootstraps_below
+            node.bootstrap_string = ""
+          end
+
+          node
+        end
+      end
+
+
+
+      tre_str = tree.newick(indent: false).gsub(/'/, '')
+
+      nexus = "#NEXUS
+begin taxa;
+dimensions ntax=#{leaves.count};
+taxlabels
+#{leaves.join("\n")}
+;
+end;
+
+begin trees;
+  tree tree_1 = [&R] #{tre_str}
+end;
+
+#{FIG}"
+
+      File.open(out_f, "w") { |f| f.puts nexus }
+    end
+  end
+end

data/lib/iroki/utils/utils.rb

@@ -0,0 +1,103 @@
+# Copyright 2016 Ryan Moore
+# Contact: moorer@udel.edu
+#
+# This file is part of Iroki.
+#
+# Iroki is free software: you can redistribute it and/or modify it
+# under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Iroki is distributed in the hope that it will be useful, but
+# WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Iroki.  If not, see <http://www.gnu.org/licenses/>.
+
+module Iroki
+  module Utils
+    include AbortIf
+    include Iroki::CoreExt::String
+
+    def leaf? tree, node
+      tree.children(node).empty?
+    end
+
+    def add_color_to_leaf_branch patterns, node, exact
+      num_matches = 0
+      color = nil
+      already_matched = false
+
+      if exact # treat patterns as string matching
+        node_s = node.to_s
+        if patterns.has_key? node_s
+          color = patterns[node_s]
+
+          return color
+        else
+          return nil
+        end
+      else
+        node_s = node.to_s
+
+        patterns.each do |pattern, this_color|
+          if node_s =~ pattern
+            abort_if already_matched,
+                     "Non specific matching for #{node_s}"
+
+            color = this_color
+            already_matched = true
+          end
+        end
+
+        return color
+      end
+    end
+
+    def get_color node
+      begin
+        node.name.match(/\[&!color="#[0-9A-Fa-f]{6}"\]/)[0]
+      rescue NoMethodError => e
+        nil
+      end
+    end
+
+    def color_nodes patterns, tree, node, exact
+      # # check if it needs color, if so set the color
+      # color = add_color_to_leaf_branch patterns, node, exact
+
+      # clean the name no matter what
+      node.name = clean_name node.name
+
+      # if its a leaf that hasnt been checked & needs color
+      if leaf?(tree, node) && !already_checked?(node.name) # && color
+        # check if it needs color, if so set the color
+
+        # NOTE: this was originally before cleaning the node name a
+        # couple lines up, does it matter that it is after?
+        color = add_color_to_leaf_branch patterns, node, exact
+
+        # add color to the name
+        node.name = node.name + color if color
+      elsif !leaf?(tree, node)
+        children = tree.children(node) # get the children
+        children_colors = []
+        children.each do |child|
+          # recurse to color the child if needed
+          color_nodes patterns, tree, child, exact
+          children_colors << get_color(child) # add color of the child
+        end
+
+        # if all the children have the same color
+        if children_colors.uniq.count == 1
+          # set the branch node to only the color name
+          node.name = children_colors[0]
+        end
+      end
+
+      return node
+    end
+  end
+end