protk 1.1.0.pre

data/lib/protk/swissprot_database.rb

@@ -0,0 +1,38 @@
+require 'rubygems'
+require 'bio'
+require 'protk/constants'
+
+# Provides fast indexed access to a swissprot database in a flat .dat file
+#
+class SwissprotDatabase 
+  
+  def initialize(env=nil,database="swissprot")
+    if ( env!=nil)
+      @genv=env
+    else
+      @genv=Constants.new
+    end
+
+    if ( database=="swissprot")
+      @db_object=Bio::FlatFileIndex.new("#{@genv.protein_database_root}/#{@genv.uniprot_sprot_annotation_database}")
+    else
+      @db_object=Bio::FlatFileIndex.new("#{@genv.protein_database_root}#{@genv.uniprot_trembl_annotation_database}")
+    end
+    
+    @db_object.always_check_consistency=false
+  end
+    
+  
+  def get_entry_for_name(name)
+    result=@db_object.get_by_id(name)
+    if result==""
+      if ( @genv!=nil)
+        @genv.log("Failed to find UniProt entry for protein named #{name} in database",:warn)
+      end
+      return nil
+    else
+      Bio::SPTR.new(result)
+    end
+  end
+  
+end

data/lib/protk/tool.rb

@@ -0,0 +1,182 @@
+#
+# This file is part of protk
+# Created by Ira Cooke 16/12/2010
+#
+# Provides common functionality used by all tools.
+#
+
+require 'ostruct'
+require 'optparse'
+require 'pathname'
+require 'protk/command_runner'
+
+class Tool
+
+  # Options set from the command-line
+  #
+  attr :options, false
+  
+  # The option parser used to parse command-line options. 
+  #
+  attr :option_parser, false
+
+  # Prefix for background jobs
+  # x = X!Tandem, o=OMSSA, p="Phenyx", m="Mascot"
+  # Can't use attr_accessor here because we want this available to subclasses
+  #
+  def jobid_prefix
+    @jobid_prefix
+  end
+
+  def jobid_prefix=(p)
+    @jobid_prefix=p
+  end
+
+  # Provides direct access to options through methods of the same name
+  #
+  def method_missing(meth, *args, &block)
+    if ( args.length==0 && block==nil)
+      @options.send meth
+    else
+      super
+    end
+  end
+  
+  
+  
+  
+  
+  # Creates an empty options object to hold commandline options
+  # Also creates an option_parser with default options common to all tools
+  #
+  def initialize(option_support={:help=>true})
+    @jobid_prefix = "x"
+    @options = OpenStruct.new
+    options.library = []
+    options.inplace = false
+    options.encoding = "utf8"
+    options.transfer_type = :auto
+    options.verbose = false
+    
+    @option_parser=OptionParser.new do |opts|
+      
+      if ( option_support[:prefix_suffix]==true)
+      
+        @options.output_prefix = ""
+        opts.on( '-b', '--output-prefix pref', 'A string to prepend to the name of output files' ) do |prefix|
+          @options.output_prefix = prefix
+        end
+
+        @options.output_suffix = ""
+        opts.on( '-e', '--output-suffix suff', 'A string to append to the name of output files' ) do |suffix|
+          @options.output_suffix = suffix
+        end
+        
+      end
+      
+      if ( option_support[:explicit_output]==true)
+        @options.explicit_output = nil
+        opts.on( '-o', '--output out', 'An explicitly named output file.' ) do |out|
+          @options.explicit_output = out
+        end
+      end
+         
+      if ( option_support[:over_write]==true)
+            
+        @options.over_write=false
+        opts.on( '-r', '--replace-output', 'Dont skip analyses for which the output file already exists' ) do  
+          @options.over_write = true
+        end
+        
+      end
+
+      if ( option_support[:background]==true)
+
+        @options.background = false
+        opts.on( '-z', '--background', 'Run jobs in the background using pbs' ) do  
+          @options.background = true
+        end
+        
+      end
+      
+      if ( option_support[:help]==true)
+       
+        opts.on( '-h', '--help', 'Display this screen' ) do
+          puts opts
+          exit
+        end
+      end
+       
+    end
+    
+  end
+  
+
+
+
+
+   # Create and return a full base path (without extension) representing the output file for this analysis
+   # Optionally provide the extension to be removed (if not provided it will be inferred)
+   #
+   def output_base_path(output_file,ext=nil)
+
+     output_path=Pathname.new(output_file)
+     throw "Error: Output directory #{output_path.dirname} does not exist" unless output_path.dirname.exist?
+     dir=output_path.dirname.realpath.to_s
+     basename=output_path.basename.to_s
+     if ( ext==nil)
+       ext=output_path.extname    
+     end
+     base_name=basename.gsub(/#{ext}$/,"")
+
+     "#{dir}/#{@options.output_prefix}#{base_name}#{@options.output_suffix}"
+   end
+
+
+
+
+   # Create a full base path (without extension) representing the input file for this analysis
+   # Optionally provide the extension to be removed (if not provided it will be inferred)
+   #
+   def input_base_path(input_file,ext=nil)
+     input_path=Pathname.new(input_file)
+     throw "Error: Input directory #{input_path.dirname} does not exist" unless input_path.dirname.exist?
+     dir=input_path.dirname.realpath.to_s
+     if ( ext==nil)
+       ext=input_path.extname    
+     end
+     base_name=input_path.basename.to_s.gsub(/#{ext}$/,"")
+     "#{dir}/#{base_name}"
+   end
+   
+   
+   
+   # Run the search tool using the given command string and global environment
+   #
+   def run(cmd,genv,job_params=nil,jobscript_path=nil,autodelete=true)
+     if ( @options.background )
+       throw "Error: Background option was selected but this host does not support background jobs" unless genv.has_pbs
+       # Send this job off to be run in a batch queuer
+       
+       cmd_runner=CommandRunner.new(genv)
+       
+ 
+       
+       cmd_runner.run_batch(cmd,job_params,jobscript_path,autodelete)
+       
+     else 
+       cmd_runner=CommandRunner.new(genv)
+       cmd_runner.run_local(cmd)
+     end
+   end
+   
+   def jobid_from_filename(filename)
+      jobid="protk"
+      jobnum_match=filename.match(/(.{1,10})\.d/)
+      if (jobnum_match!=nil)
+        jobid="#{self.jobid_prefix}#{jobnum_match[1]}"
+      end
+      return jobid
+    end
+  
+end

data/lib/protk/xtandem_defaults.rb

@@ -0,0 +1,11 @@
+require 'libxml'
+include LibXML
+
+class XTandemDefaults
+	attr :path
+	attr :taxonomy_path
+	def initialize
+		@path="#{File.dirname(__FILE__)}/data/tandem_params.xml"
+		@taxonomy_path="#{File.dirname(__FILE__)}/data/taxonomy_template.xml"
+	end
+end

data/lib/protk.rb

@@ -0,0 +1,18 @@
+require 'protk/tool.rb'
+require 'protk/swissprot_database.rb'
+require 'protk/spreadsheet_extensions.rb'
+require 'protk/search_tool.rb'
+require 'protk/protxml.rb'
+require 'protk/protein_annotator.rb'
+require 'protk/prophet_tool.rb'
+require 'protk/plasmodb.rb'
+require 'protk/omssa_util.rb'
+require 'protk/mascot_util.rb'
+require 'protk/manage_db_tool.rb'
+require 'protk/galaxy_util.rb'
+require 'protk/galaxy_stager.rb'
+require 'protk/eupathdb_gene_information_table.rb'
+require 'protk/constants.rb'
+require 'protk/command_runner.rb'
+require 'protk/biotools_excel_converter.rb'
+require 'protk/bio_sptr_extensions.rb'

metadata

@@ -0,0 +1,256 @@
+--- !ruby/object:Gem::Specification
+name: protk
+version: !ruby/object:Gem::Version
+  version: 1.1.0.pre
+  prerelease: 6
+platform: ruby
+authors:
+- Ira Cooke
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2012-10-19 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: ftools
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.0.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.0.0
+- !ruby/object:Gem::Dependency
+  name: open4
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 1.3.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 1.3.0
+- !ruby/object:Gem::Dependency
+  name: bio
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 1.4.3
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 1.4.3
+- !ruby/object:Gem::Dependency
+  name: rest-client
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 1.6.7
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 1.6.7
+- !ruby/object:Gem::Dependency
+  name: net-ftp-list
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 3.2.5
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 3.2.5
+- !ruby/object:Gem::Dependency
+  name: spreadsheet
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.7.4
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.7.4
+- !ruby/object:Gem::Dependency
+  name: libxml-ruby
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 2.3.3
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 2.3.3
+- !ruby/object:Gem::Dependency
+  name: rspec
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '2.5'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '2.5'
+description: A bunch of tools for proteomics
+email: iracooke@gmail.com
+executables:
+- protk_setup.rb
+- manage_db.rb
+- tandem_search.rb
+- mascot_search.rb
+- omssa_search.rb
+- msgfplus_search.rb
+- big_search.rb
+- mascot_to_pepxml.rb
+- file_convert.rb
+- make_decoy.rb
+- correct_omssa_retention_times.rb
+- repair_run_summary.rb
+- peptide_prophet.rb
+- interprophet.rb
+- protein_prophet.rb
+- pepxml_to_table.rb
+- xls_to_table.rb
+- annotate_ids.rb
+- unimod_to_loc.rb
+- generate_omssa_loc.rb
+extensions:
+- ext/protk/extconf.rb
+extra_rdoc_files: []
+files:
+- lib/protk/big_search_tool.rb
+- lib/protk/bio_sptr_extensions.rb
+- lib/protk/biotools_excel_converter.rb
+- lib/protk/command_runner.rb
+- lib/protk/constants.rb
+- lib/protk/eupathdb_gene_information_table.rb
+- lib/protk/galaxy_stager.rb
+- lib/protk/galaxy_util.rb
+- lib/protk/manage_db_tool.rb
+- lib/protk/mascot_util.rb
+- lib/protk/omssa_util.rb
+- lib/protk/plasmodb.rb
+- lib/protk/prophet_tool.rb
+- lib/protk/protein_annotator.rb
+- lib/protk/protxml.rb
+- lib/protk/randomize.rb
+- lib/protk/search_tool.rb
+- lib/protk/setup_tool.rb
+- lib/protk/spreadsheet_extensions.rb
+- lib/protk/swissprot_database.rb
+- lib/protk/tool.rb
+- lib/protk/xtandem_defaults.rb
+- lib/protk.rb
+- lib/protk/big_search_rakefile.rake
+- lib/protk/manage_db_rakefile.rake
+- lib/protk/setup_rakefile.rake
+- bin/annotate_ids.rb
+- bin/big_search.rb
+- bin/correct_omssa_retention_times.rb
+- bin/feature_finder.rb
+- bin/file_convert.rb
+- bin/generate_omssa_loc.rb
+- bin/interprophet.rb
+- bin/make_decoy.rb
+- bin/manage_db.rb
+- bin/mascot_search.rb
+- bin/mascot_to_pepxml.rb
+- bin/msgfplus_search.rb
+- bin/omssa_search.rb
+- bin/peptide_prophet.rb
+- bin/pepxml_to_table.rb
+- bin/protein_prophet.rb
+- bin/protk_setup.rb
+- bin/repair_run_summary.rb
+- bin/tandem_search.rb
+- bin/template_search.rb
+- bin/unimod_to_loc.rb
+- bin/xls_to_table.rb
+- README.md
+- lib/protk/data/apt-get_packages.yaml
+- lib/protk/data/brew_packages.yaml
+- lib/protk/data/default_config.yml
+- lib/protk/data/FeatureFinderCentroided.ini
+- lib/protk/data/predefined_db.crap.yaml
+- lib/protk/data/predefined_db.sphuman.yaml
+- lib/protk/data/predefined_db.swissprot_annotation.yaml
+- lib/protk/data/predefined_db.swissprot_fasta_annotation.yaml
+- lib/protk/data/tandem_params.xml
+- lib/protk/data/taxonomy_template.xml
+- lib/protk/data/unimod.xml
+- ext/protk/protk.c
+- ext/protk/extconf.rb
+homepage: http://rubygems.org/gems/protk
+licenses: []
+post_install_message: Now run protk_setup.rb to install third party tools and manage_db.rb
+  to install databases.
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>'
+    - !ruby/object:Gem::Version
+      version: 1.3.1
+requirements: []
+rubyforge_project: 
+rubygems_version: 1.8.24
+signing_key: 
+specification_version: 3
+summary: Proteomics Toolkit
+test_files: []