Class: BioDSL::Usearch::IO

Inherits:
Filesys
  • Object
show all
Defined in:
lib/BioDSL/usearch.rb

Overview

Class for Usearch IO.

Instance Attribute Summary

Attributes inherited from Filesys

#io

Instance Method Summary collapse

Methods inherited from Filesys

#close, #eof?, #gets, #initialize, open, #puts, #read, tmpfile, which, #write

Constructor Details

This class inherits a constructor from BioDSL::Filesys

Instance Method Details

#each(format = :uc) ⇒ Object

Parse a given type of Uclust format and yield the result.

Parameters:

  • format (Symbol) (defaults to: :uc) —

    Format type to parse.



266
267
268
269
270
271
272
 
# File 'lib/BioDSL/usearch.rb', line 266

def each(format = :uc)
  case format
  when :uc then each_uc { |e| yield e }
  else
    fail UsearchError, "Unknown iterator format: #{format}"
  end
end

#each_uc {|Hash| ... } ⇒ Object

Parse each UC type record and yield the result.

Yields:

  • (Hash) —

    BioDSL record with UC result.



279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
 
# File 'lib/BioDSL/usearch.rb', line 279

def each_uc
  @io.each do |line|
    fields = line.chomp.split("\t")
    record = {TYPE:    fields[0],
              CLUSTER: fields[1].to_i}

    case fields[0]
    when 'C' then record[:CLUSTER_SIZE] = fields[2].to_i
    else          record[:SEQ_LEN]      = fields[2].to_i
    end

    record[:IDENT]  = fields[3].to_f if fields[0] == 'H'
    record[:STRAND] = fields[4]
    record[:CIGAR]  = fields[7]
    record[:Q_ID]   = fields[8]
    record[:S_ID]   = fields[9] if fields[0] == 'H'

    yield record
  end
end