sequenceserver 0.8.9 → 1.0.0.pre.1

data/lib/sequenceserver/customisation.rb

@@ -1,60 +0,0 @@
-module SequenceServer
-  module Customisation
-    ## When not commented out, this method is used to take a
-    ## sequence ID, and return a hyperlink that
-    ## replaces the hit in the BLAST output.
-    ##
-    ## Return the hyperlink to link to, or nil
-    ## to not not include a hyperlink.
-    ##
-    ## When this method
-    ## is commented out, the default link is used. The default
-    ## is a link to the full sequence of
-    ## the hit is displayed (if makeblastdb has been run with
-    ## -parse_seqids), or no link at all otherwise.
-    # def construct_custom_sequence_hyperlink(options)
-    #   ## Example:
-    #   ## sequence_id comes in like "psu|MAL13P1.200 | organism=Plasmodium_falciparum_3D7 | product=mitochondrial"
-    #   ## output: "http://apiloc.bio21.unimelb.edu.au/apiloc/gene/MAL13P1.200"
-    #   matches = options[:sequence_id].match(/^\s*psu\|(\S+) /)
-    #   if matches #if the sequence_id conforms to our expectations
-    #     # All is good. Return the hyperlink.
-    #     return "http://apiloc.bio21.unimelb.edu.au/apiloc/gene/#{matches[1]}"
-    #   else
-    #     # Parsing the sequence_id didn't work. Don't include a hyperlink for this
-    #     # sequence_id, but log that there has been a problem.
-    #     settings.log.warn "Unable to parse sequence id `#{options[:sequence_id]}'"
-    #     # Return nil so no hyperlink is generated.
-    #     return nil
-    #   end
-    # end
-
-    ## Much like construct_custom_sequence_hyperlink, except
-    ## instead of just a hyperlink being defined, the whole
-    ## line as it appears in the blast results is generated.
-    ##
-    ## This is a therefore more flexible setup than is possible
-    ## with construct_custom_sequence_hyperlink, because doing
-    ## things such as adding two hyperlinks for the one hit
-    ## are possible.
-    ##
-    ## When this method is commented out, the behaviour is that
-    ## the construct_custom_sequence_hyperlink method is used,
-    ## or failing that the default method of that is used.
-    # def construct_custom_sequence_hyperlinking_line(options)
-    #   matches = options[:sequence_id].match(/^\s*psu\|(\S+) /)
-    #   if matches #if the sequence_id conforms to our expectations
-    #     # All is good. Return the hyperlink.
-    #     link1 = "http://apiloc.bio21.unimelb.edu.au/apiloc/gene/#{matches[1]}"
-    #     link2 = "http://google.com/?q=#{matches[1]}"
-    #     return "<a href='#{link1}'>ApiLoc page</a>, <a href='#{link2}'>Google search</a>"
-    #   else
-    #     # Parsing the sequence_id didn't work. Don't include a hyperlink for this
-    #     # sequence_id, but log that there has been a problem.
-    #     settings.log.warn "Unable to parse sequence id `#{options[:sequence_id]}'"
-    #     # Return nil so no hyperlink is generated.
-    #     return nil
-    #   end
-    # end
-  end
-end

data/lib/sequenceserver/database_formatter.rb

@@ -1,190 +0,0 @@
-# copyright yannick . wurm at unil . ch
-# Finds files, reads first char. if its '>', read 500 lines. Guess sequence type, ask user for title to format as blast database.
-
-# TODO: bring it under SequenceServer namespace
-# TODO: move the file to a 'command/' sub-directory (probably makes more sense if we have several subcommands)
-# TODO: needs more love (read refactoring) overall
-
-require 'ptools' # for File.binary?(file)
-require 'find'
-require 'logger'
-require 'optparse'
-require 'sequenceserver'
-require 'sequenceserver/helpers.rb'
-require 'sequenceserver/sequencehelpers.rb'
-
-LOG = Logger.new(STDOUT)
-LOG.level = Logger::INFO
-
-class DatabaseFormatter
-    include SequenceServer
-	include Helpers
-    include SystemHelpers
-    include SequenceHelpers
-    
-    attr_accessor :db_path
-
-    def initialize(db_path = nil)
-      @app = SequenceServer::App
-      @app.config = @app.parse_config
-      @app.binaries = @app.scan_blast_executables(@app.bin).freeze
-
-      @db_path = (db_path or @app.database)
-    end
-
-    def format_databases
-      unless File.directory?(db_path)
-        LOG.fatal("Database directory #{db_path} not found. See './database_formatter --help' for instructions.")
-        exit
-      end
-
-        formatted_dbs = %x|#{@app.binaries['blastdbcmd']} -recursive -list #{db_path} -list_outfmt "%f" 2>&1|.split("\n")
-        commands = []
-        Find.find(db_path) do |file|
-            LOG.debug("Assessing file #{file}..")
-            if File.directory?(file)
-              LOG.debug("Ignoring file #{file} since it is a directory")
-              next
-            end
-            if formatted_dbs.include?(file)
-              LOG.debug("Ignoring file #{file} since it is already a blast database")
-              next
-            end
-            if File.binary?(file)
-              LOG.debug("Ignoring file #{file} since it is a binary file, not plaintext as FASTA files are")
-              next
-            end
-
-                if probably_fasta?(file)
-                    LOG.info("Found #{file}")
-                    ## guess whether protein or nucleotide based on first 500 lines
-                    first_lines = ''
-                    File.open(file, 'r') do |file_stream|
-                        file_stream.each do |line|
-                            first_lines += line
-                            break if file_stream.lineno == 500
-                        end
-                    end
-                    begin
-                        sequence_type = type_of_sequences(first_lines) # returns :protein or :nucleotide
-                    rescue
-                        LOG.warn("Unable to guess sequence type for #{file}. Skipping") 
-                    end
-                    if [ :protein, :nucleotide ].include?(sequence_type)
-                        command = ask_make_db_command(file, sequence_type)
-                        unless command.nil?
-                            commands.push(command)
-                        end
-                    else 
-                        LOG.warn("Unable to guess sequence type for #{file}. Skipping") 
-                    end
-                else
-                  LOG.debug("Ignoring file #{file} since it was not judged to be a FASTA file.")
-                end
-        end
-        LOG.info("Will now create DBs")
-        if commands.empty?
-          puts "", "#{db_path} does not contain any unformatted database."
-          exit
-        end
-        commands.each do |command|
-			LOG.info("Will run: " + command.to_s)
-            system(command)
-        end
-        LOG.info("Done formatting databases. ")
-        db_table(db_path)
-    end
-
-    def db_table(db_path)
-        LOG.info("Summary of formatted blast databases:\n")
-        output = %x|#{@app.binaries['blastdbcmd']} -recursive -list #{db_path} -list_outfmt "%p %f %t" &2>1 |
-        LOG.info(output)
-    end
-
-    def probably_fasta?(file)
-		return FALSE if File.zero?(file)
-        File.open(file, 'r') do |file_stream|
-            first_line = file_stream.readline
-            if first_line.slice(0,1) == '>'
-                return TRUE
-            else 
-                return FALSE
-            end
-        end
-    end
-
-
-    # returns command than needs to be run to make db
-    def ask_make_db_command(file, type)
-        LOG.info("FASTA file: #{file}")
-        LOG.info("Fasta type: " + type.to_s)
-        
-        response = ''
-        until response.match(/^[yn]$/i) do
-            LOG.info("Proceed? [y/n]: ")
-            response = STDIN.gets.chomp
-        end
-
-        if response.match(/y/i)
-            LOG.info("Enter a database title (or will use '#{File.basename(file)}'")
-            title = STDIN.gets.chomp
-			title.gsub!('"', "'")
-            title = File.basename(file)  if title.empty?
-            
-            return make_db_command(file,type,title)
-        end
-    end
-
-    def make_db_command(file,type, title)
-        LOG.info("Will make #{type.to_s} database from #{file} with #{title}")
-        command = %|#{@app.binaries['makeblastdb']} -in #{file} -dbtype #{ type.to_s.slice(0,4)} -title "#{title}" -parse_seqids|
-        LOG.info("Returning: #{command}")
-        return(command)
-    end
-end
-
-OptionParser.new do |opts|
-  opts.banner =<<BANNER
-
-SUMMARY
-
-  prepare BLAST databases for SequenceServer
-
-USAGE
-
-  sequenceserver format-databases [--verbose] [blast_database_directory]
-
-  Example:
-
-    $ sequenceserver format-databases ~/db  # explicitly specify a database directory
-    $ sequenceserver format-databases      # use the database directory in config.yml
-
-DESCRIPTION
-
-  Recursively scan the given 'blast_database_directory' for BLAST databases and
-  formats them for use with SequenceServer.
-
-  It automagically detects the database type, and ignores non-db files and
-  pre-formatted databases. The 'parse_seqids' makeblastdb options is used.
-
-  'blast_database_directory' can be passed as a command line parameter or
-  through a configuration file by setting the 'database' key (the same option
-  used by SequenceServer). Configuration file will be checked only if the
-  command line parameter is missing.
-
-OPTIONS
-
-BANNER
-
-  opts.on_tail('-h', '--help', 'Show this message') do
-    puts opts
-    exit
-  end
-
-  opts.on('-v', '--verbose', 'Print lots of output') do
-    LOG.level = Logger::DEBUG
-  end
-end.parse!
-
-app = DatabaseFormatter.new(ARGV[0])
-app.format_databases

data/lib/sequenceserver/helpers.rb

@@ -1,136 +0,0 @@
-require 'sequenceserver/database'
-
-module SequenceServer
-  module Helpers
-    module SystemHelpers
-      # Scan the given directory for blast executables. Passing `nil` scans the
-      # system `PATH`.
-      # ---
-      # Arguments:
-      # * bin(String) - absolute path to the directory containing blast binaries
-      # ---
-      # Returns:
-      # * a hash of blast methods, and their corresponding absolute path
-      # ---
-      # Raises:
-      # * IOError - if the executables can't be found
-      #
-      #   > scan_blast_executables('/home/yeban/bin')
-      #   =>  { "blastx"=>"/home/yeban/bin/blastx",
-      #         "blastn"=>"/home/yeban/bin/blastn",
-      #         ...
-      #       }
-      def scan_blast_executables(bin)
-        if bin and not File.directory?(bin)
-          raise IOError, "Could not find '#{bin}' defined in config.yml."
-        end
-
-        binaries = {}
-        %w|blastn blastp blastx tblastn tblastx blastdbcmd makeblastdb blast_formatter|.each do |method|
-          path = File.join(bin, method) rescue method
-          if command?(path)
-            binaries[method] = path
-          else
-            blasturl = 'http://www.ncbi.nlm.nih.gov/blast/Blast.cgi?CMD=Web&PAGE_TYPE=BlastDocs&DOC_TYPE=Download'
-            raise IOError, "Could not find blast binaries." +
-            "\n\nYou may need to download BLAST+ from #{blasturl}." +
-            " And/or edit #{settings.config_file} to indicate the location of BLAST+ binaries."
-          end
-        end
-
-        #LOG.info("Config bin dir:          #{bin}")
-        binaries
-      end
-
-      # Scan the given directory (including subdirectory) for blast databases.
-      # ---
-      # Arguments:
-      # * db_root(String) - absolute path to the blast databases
-      # ---
-      # Returns:
-      # * a hash of sorted blast databases grouped by database type:
-      # protein, or nucleotide
-      # ---
-      # Raises:
-      # * IOError - if no database can be found
-      #
-      #   > scan_blast_db('/home/yeban/blast_db')
-      #   => { "protein" => [], "nucleotide" => [] }
-      def scan_blast_db(db_root, blastdbcmd = 'blastdbcmd')
-        raise IOError, "Database directory doesn't exist: #{db_root}" unless File.directory?( db_root )
-
-        find_dbs_command = %|#{blastdbcmd} -recursive -list #{db_root} -list_outfmt "%p %f %t" 2>&1|
-
-        begin
-          db_list = %x|#{find_dbs_command}|
-          if db_list.empty?
-            raise IOError, "No formatted blast databases found in '#{ db_root }'."
-          end
-        rescue => e
-          puts '', e.to_s
-
-          print "Do you want to format your blast databases now? [Y/n]: "
-          choice = gets.chomp[0,1].downcase
-
-          unless choice == 'n'
-            database_formatter = File.join(settings.root, 'lib', 'sequenceserver', 'database_formatter.rb')
-            require database_formatter
-            retry
-          else
-            raise # let the caller decide what to do if database discovery fails
-          end
-        end
-
-        if db_list.match(/BLAST Database error/)
-          raise IOError, "Error parsing blast databases.\n" + "Tried: '#{find_dbs_command}'\n"+
-            "It crashed with the following error: '#{db_list}'\n" +
-            "Try reformatting databases using makeblastdb.\n"
-        end
-
-        db = {}
-
-        db_list.each_line do |line|
-          next if line.empty?  # required for BLAST+ 2.2.22
-          type, name, *title =  line.split(' ') 
-          type = type.downcase.intern
-          name = name.freeze
-          title = title.join(' ').freeze
-
-          # skip past all but alias file of a NCBI multi-part BLAST database
-          if multipart_database_name?(name)
-            log.info(%|Found a multi-part database volume at #{name} - ignoring it.|)
-            next
-          end
-
-          #LOG.info("Found #{type} database: #{title} at #{name}")
-          database = Database.new(name, title, type)
-          db[database.hash] = database
-        end
-
-        db 
-      end
-
-      private
-
-      # check if the given command exists and is executable
-      # returns True if all is good.
-      def command?(command)
-        system("which #{command} > /dev/null 2>&1")
-      end
-
-      # Returns true if the database name appears to be a multi-part database name.
-      #
-      # e.g.
-      # /home/ben/pd.ben/sequenceserver/db/nr.00 => yes
-      # /home/ben/pd.ben/sequenceserver/db/nr => no
-      # /home/ben/pd.ben/sequenceserver/db/img3.5.finished.faa.01 => yes
-      def multipart_database_name?(db_name)
-        !(db_name.match(/.+\/\S+\d{2}$/).nil?)
-      end
-    end
-
-    def self.included(klass)
-      klass.extend SystemHelpers
-    end
-  end
-end

data/lib/sequenceserver/sequencehelpers.rb

@@ -1,93 +0,0 @@
-module SequenceServer
-  # Module to collect some sequence-related helper functions
-  module SequenceHelpers
-
-    # copied from bioruby's Bio::Sequence
-    # returns a Hash. Eg: composition("asdfasdfffffasdf")
-    #                      => {"a"=>3, "d"=>3, "f"=>7, "s"=>3}
-    def composition(sequence_string)
-      count = Hash.new(0)
-      sequence_string.scan(/./) do |x|
-        count[x] += 1
-      end
-      return count
-    end
-
-    # Strips all non-letter characters. guestimates sequence based on that.
-    # If less than 10 useable characters... returns nil
-    # If more than 90% ACGTU returns :nucleotide. else returns :protein
-    def guess_sequence_type(sequence_string)
-      cleaned_sequence = sequence_string.gsub(/[^A-Z]/i, '')  # removing non-letter characters
-      cleaned_sequence.gsub!(/[NX]/i, '')                     # removing ambiguous  characters
-
-      return nil if cleaned_sequence.length < 10 # conservative
-
-      composition = composition(cleaned_sequence)       
-      composition_NAs    = composition.select { |character, count|character.match(/[ACGTU]/i) } # only putative NAs
-      putative_NA_counts = composition_NAs.collect { |key_value_array| key_value_array[1] }     # only count, not char
-      putative_NA_sum    = putative_NA_counts.inject { |sum, n| sum + n }                       # count of all putative NA
-      putative_NA_sum    = 0 if putative_NA_sum.nil?
-
-      if putative_NA_sum > (0.9 * cleaned_sequence.length)
-        return :nucleotide
-      else
-        return :protein
-      end
-    end
-
-    # splits input at putative fasta definition lines (like ">adsfadsf"), guesses sequence type for each sequence. 
-    # if not enough sequence to determine, returns nil.
-    # if 2 kinds of sequence mixed together, raises ArgumentError
-    # otherwise, returns :nucleotide or :protein
-    def type_of_sequences(fasta_format_string)
-      # the first sequence does not need to have a fasta definition line
-      sequences = fasta_format_string.split(/^>.*$/).delete_if { |seq| seq.empty? }
-
-      # get all sequence types
-      sequence_types = sequences.collect { |seq|  guess_sequence_type(seq) }.uniq.compact
-
-      return nil if sequence_types.empty?
-
-      if sequence_types.length == 1
-        return sequence_types.first # there is only one (but yes its an array)
-      else 
-        raise ArgumentError, "Insufficient info to determine sequence type. Cleaned queries are: #{ sequences.to_s }"
-      end
-    end
-
-    def sequence_from_blastdb(ids, db)  # helpful when displaying parsed blast results
-      # we know how to handle an Array of ids
-      ids = ids.join(',') if ids.is_a? Array
-
-      # we don't know what to do if the arguments ain't String
-      raise TypeError unless ids.is_a? String and db.is_a? String
-
-      # query now!
-      #
-      # If `blastdbcmd` throws error, we assume sequence not found.
-      blastdbcmd = settings.binaries['blastdbcmd']
-      %x|#{blastdbcmd} -db #{db} -entry '#{ids}' 2> /dev/null|
-    end
-
-    # Given a sequence_id and databases, apply the default (standard)
-    # way to convert a sequence_id into a hyperlink, so that the
-    # blast results include hyperlinks.
-    def construct_standard_sequence_hyperlink(options)
-      if options[:sequence_id].match(/^[^ ]/) #if there is a space right after the '>', makeblastdb was run without -parse_seqids
-        # By default, add a link to a fasta file of the sequence (if makeblastdb was called with -parse_seqids)
-
-        sid = options[:sequence_id].gsub(/<\/?[^>]*>/, '') # strip html
-        cid = sid[/^(\S+)\s*.*/, 1]                        # get id part
-        id  = cid.include?('|') ? cid.split('|')[1] : cid.split('|')[0]
-        @all_retrievable_ids ||= []
-        @all_retrievable_ids.push(id)
-
-        link = "/get_sequence/?id=#{id}&db=#{options[:databases].join(' ')}" # several dbs... separate by ' '
-        return link
-      else
-        # do nothing - link == nil means no link will be incorporated
-        return nil
-      end
-    end
-  end
-end