viral_seq 1.0.7 → 1.0.12

data/lib/viral_seq/tcs_json.rb

@@ -0,0 +1,178 @@
+module ViralSeq
+  class TcsJson
+    class << self
+
+      def generate
+        puts '-'*58
+        puts '| JSON Parameter Generator for ' + "TCS #{ViralSeq::TCS_VERSION}".red.bold + " by " + "Shuntai Zhou".blue.bold + ' |'
+        puts '-'*58 + "\n"
+
+        param = {}
+
+        puts 'Enter the path to the directory that contains the MiSeq pair-end R1 and R2 .fastq or .fastq.gz file'
+        print '> '
+        param[:raw_sequence_dir] = gets.chomp.rstrip
+
+        puts 'Enter the estimated platform error rate (for TCS cut-off calculation), default as ' + '0.02'.red.bold
+        print '> '
+        input_error = gets.chomp.rstrip.to_f
+        if input_error == 0.0
+          param[:platform_error_rate] = 0.02
+        else
+          param[:platform_error_rate] = input_error
+        end
+
+        param[:primer_pairs] = []
+
+        loop do
+          data = {}
+          puts "Enter the name for the sequenced region: "
+          print '> '
+          data[:region] = gets.chomp.rstrip
+
+          puts "Enter the #{"cDNA".red.bold} primer sequence: "
+          print '> '
+          data[:cdna] = gets.chomp.rstrip
+
+          puts "Enter the #{"forward".blue.bold} primer sequence: "
+          print '> '
+          data[:forward] = gets.chomp.rstrip
+
+          puts "Enter supermajority cut-off (0.5 - 1.0). Default Simple Majority"
+          print '> '
+          mj = gets.chomp.rstrip.to_f
+          if (0.5..1.0).include?(mj)
+            data[:majority] = mj
+          else
+            data[:majority] = 0
+          end
+
+          print "Need end-join? Y/N \n> "
+          ej = gets.chomp.rstrip
+          if ej =~ /y|yes/i
+            data[:end_join] = true
+
+            print "End-join option? Choose from (1-4):\n
+            1: simple join, no overlap
+            2: known overlap \n
+            3: unknow overlap, use sample consensus to determine overlap, all sequence pairs have same overlap\n
+            4: unknow overlap, determine overlap by individual sequence pairs, sequence pairs can have different overlap\n
+            > "
+            ej_option = gets.chomp.rstrip
+            while ![1,2,3,4].include?(ej_option.to_i)
+              puts "Entered end-join option #{ej_option.red.bold} not valid (choose 1-4), try again"
+              ej_option = gets.chomp.rstrip.to_i
+            end
+            case ej_option.to_i
+            when 1
+              data[:end_join_option] = 1
+              data[:overlap] = 0
+            when 2
+              data[:end_join_option] = 1
+              print "overlap bases: \n> "
+              ol = gets.chomp.rstrip.to_i
+              data[:overlap] = ol
+            when 3
+              data[:end_join_option] = 3
+            when 4
+              data[:end_join_option] = 4
+            end
+
+            print "Need QC for TCS? (support for HIV-1 and SIV)? Y/N \n> "
+            qc = gets.chomp.rstrip
+            if qc =~ /y|yes/i
+              data[:TCS_QC] = true
+
+              data[:ref_genome] = get_ref
+
+              print "reference 5'end ref position or posiiton range, 0 if no need to match this end \n> "
+              data[:ref_start] = gets.chomp.rstrip.to_i
+
+              print "reference 3'end ref position or posiiton range: 0 if no need to match this end \n> "
+              data[:ref_end] = gets.chomp.rstrip.to_i
+
+              print "allow indels? (default as yes) Y/N \n> "
+              indel = gets.chomp.rstrip
+              if indel =~ /n|no/i
+                data[:indel] = false
+              else
+                data[:indel] = true
+              end
+            else
+              data[:TCS_QC] = false
+            end
+
+            print "Need trimming to a reference genome? Y/N \n> "
+            trim_option = gets.chomp.rstrip
+            if trim_option =~ /y|yes/i
+              data[:trim] = true
+              data[:trim_ref] = get_ref
+
+              print "reference 5'end ref position \n> "
+              data[:trim_ref_start] = gets.chomp.rstrip.to_i
+
+              print "reference 3'end ref position \n> "
+              data[:trim_ref_end] = gets.chomp.rstrip.to_i
+
+            else
+              data[:trim] = false
+            end
+
+          else
+            data[:end_join] = false
+          end
+
+          param[:primer_pairs] << data
+          print "Do you wish to conintue? Y/N \n> "
+          continue_sig = gets.chomp.rstrip
+          break unless continue_sig =~ /y|yes/i
+
+        end
+
+        puts "\nYour JSON string is:"
+        puts JSON.pretty_generate(param)
+
+        print "\nDo you wish to save it as a file? Y/N \n> "
+        save_option = gets.chomp.rstrip
+
+        if save_option =~ /y|yes/i
+          print "Path to save JSON file:\n> "
+          path = gets.chomp.rstrip
+          File.open(path, 'w') {|f| f.puts JSON.pretty_generate(param)}
+        end
+
+        print "\nDo you wish to execute tcs pipeline with the input params now? Y/N \n> "
+
+        rsp = gets.chomp.rstrip
+        if rsp =~ /y/i
+          return param
+        else
+          abort "Params json file generated. You can execute tcs pipeline using `tcs -p [params.json]`"
+        end
+
+      end
+
+      private
+      def get_ref
+        puts "Choose reference genome (1-3):"
+        puts "1. HIV-1 HXB2".red.bold
+        puts "2. HIV-1 NL4-3".blue.bold
+        puts "3. SIV MAC239".magenta.bold
+        print "> "
+        ref_option = gets.chomp.rstrip
+        while ![1,2,3].include?(ref_option.to_i)
+          print "Entered end-join option #{ref_option.to_s.red.bold} not valid (choose 1-3), try again\n> "
+          ref_option = gets.chomp.rstrip.to_i
+        end
+        ref = case ref_option.to_i
+              when 1
+                :HXB2
+              when 2
+                :NL43
+              when 3
+                :MAC239
+              end
+      end
+    end
+  end # end TcsJson
+end # end main module

data/lib/viral_seq/version.rb

@@ -2,5 +2,6 @@
 # version info and histroy
 
 module ViralSeq
-  VERSION = "1.0.7"
+  VERSION = "1.0.12"
+  TCS_VERSION = "2.1.1"
 end

data/viral_seq.gemspec

@@ -26,7 +26,7 @@ Gem::Specification.new do |spec|
   spec.post_install_message = "Thanks for installing!"
 
   spec.add_development_dependency "bundler", "~> 2.0"
-  spec.add_development_dependency "rake", "~> 10.0"
+  spec.add_development_dependency "rake", "~> 13.0"
   spec.add_development_dependency "rspec", "~> 3.0"
 
   # muscle_bio gem required

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: viral_seq
 version: !ruby/object:Gem::Version
-  version: 1.0.7
+  version: 1.0.12
 platform: ruby
 authors:
 - Shuntai Zhou
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-01-28 00:00:00.000000000 Z
+date: 2021-03-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -31,14 +31,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '10.0'
+        version: '13.0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '10.0'
+        version: '13.0'
 - !ruby/object:Gem::Dependency
   name: rspec
   requirement: !ruby/object:Gem::Requirement
@@ -89,6 +89,7 @@ email:
 - clarkmu@gmail.com
 executables:
 - locator
+- tcs
 extensions: []
 extra_rdoc_files: []
 files:
@@ -102,6 +103,7 @@ files:
 - README.md
 - Rakefile
 - bin/locator
+- bin/tcs
 - lib/viral_seq.rb
 - lib/viral_seq/constant.rb
 - lib/viral_seq/enumerable.rb
@@ -112,10 +114,13 @@ files:
 - lib/viral_seq/pid.rb
 - lib/viral_seq/ref_seq.rb
 - lib/viral_seq/rubystats.rb
+- lib/viral_seq/sdrm.rb
 - lib/viral_seq/seq_hash.rb
 - lib/viral_seq/seq_hash_pair.rb
 - lib/viral_seq/sequence.rb
 - lib/viral_seq/string.rb
+- lib/viral_seq/tcs_core.rb
+- lib/viral_seq/tcs_json.rb
 - lib/viral_seq/version.rb
 - viral_seq.gemspec
 homepage: https://github.com/ViralSeq/viral_seq
@@ -138,7 +143,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements:
 - R required for some functions
-rubygems_version: 3.1.2
+rubygems_version: 3.2.2
 signing_key: 
 specification_version: 4
 summary: A Ruby Gem containing bioinformatics tools for processing viral NGS data.