Class: Bio::FlatFileIndex::Flat_1::FlatMappingFile

Inherits:
Object
  • Object
show all
Defined in:
lib/bio/io/flatfile/index.rb

Overview

FlatMappingFile class.

Internal use only.

Constant Summary collapse

@@recsize_width = 
4
@@recsize_regex = 
/\A\d{4}\z/

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(filename, mode = 'rb') ⇒ FlatMappingFile

Returns a new instance of FlatMappingFile.



734
735
736
737
738
739
740
741
 
# File 'lib/bio/io/flatfile/index.rb', line 734

def initialize(filename, mode = 'rb')
  @filename = filename
  @mode = mode
  @file = nil
  #@file = File.open(filename, mode)
  @record_size = nil
  @records = nil
end

Instance Attribute Details

#filename ⇒ Object (readonly)

Returns the value of attribute filename.



743
744
745
 
# File 'lib/bio/io/flatfile/index.rb', line 743

def filename
  @filename
end

#mode ⇒ Object

Returns the value of attribute mode.



742
743
744
 
# File 'lib/bio/io/flatfile/index.rb', line 742

def mode
  @mode
end

Class Method Details

.external_merge_proc(sort_program = [ '/usr/bin/env', 'LC_ALL=C', '/usr/bin/sort' ]) ⇒ Object 



923
924
925
926
927
928
929
930
931
 
# File 'lib/bio/io/flatfile/index.rb', line 923

def self.external_merge_proc(sort_program =  [ '/usr/bin/env', 
                                               'LC_ALL=C',
                                               '/usr/bin/sort' ])
  Proc.new do |out, in1, *files|
    # files (and in1) must be sorted
    cmd = sort_program + [ '-m', '-o', out, in1, *files ]
    system(*cmd)
  end
end

.external_merge_sort_proc(sort_program = [ '/usr/bin/env', 'LC_ALL=C', '/usr/bin/sort' ]) ⇒ Object 



900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
 
# File 'lib/bio/io/flatfile/index.rb', line 900

def self.external_merge_sort_proc(sort_program = [ '/usr/bin/env', 
                                                   'LC_ALL=C',
                                                   '/usr/bin/sort' ])
  Proc.new do |out, in1, *files|
    # (in1 may be sorted)
    tf_all = []
    tfn_all = []
    files.each do |fn|
      tf = Tempfile.open('sort')
      tf.close(false)
      cmd = sort_program + [ '-o', tf.path, fn ]
      system(*cmd)
      tf_all << tf
      tfn_all << tf.path
    end
    cmd_fin = sort_program + [ '-m', '-o', out, in1, *tfn_all ]
    system(*cmd_fin)
    tf_all.each do |tf|
      tf.close(true)
    end
  end
end

.external_sort_proc(sort_program = [ '/usr/bin/env', 'LC_ALL=C', '/usr/bin/sort' ]) ⇒ Object 



891
892
893
894
895
896
897
898
 
# File 'lib/bio/io/flatfile/index.rb', line 891

def self.external_sort_proc(sort_program = [ '/usr/bin/env', 
                                             'LC_ALL=C',
                                             '/usr/bin/sort' ])
  Proc.new do |out, in1, *files|
    cmd = sort_program + [ '-o', out, in1, *files ]
    system(*cmd)
  end
end

.internal_sort_proc ⇒ Object 



933
934
935
936
937
938
939
940
941
942
943
944
945
946
 
# File 'lib/bio/io/flatfile/index.rb', line 933

def self.internal_sort_proc
  Proc.new do |out, in1, *files|
    a = IO.readlines(in1)
    files.each do |fn|
      IO.foreach(fn) do |x|
        a << x
      end
    end
    a.sort!
    of = File.open(out, 'w')
    a.each { |x| of << x }
    of.close
  end
end

.open(*arg) ⇒ Object 



730
731
732
 
# File 'lib/bio/io/flatfile/index.rb', line 730

def self.open(*arg)
  self.new(*arg)
end

Instance Method Details

#add_record(str) ⇒ Object 



806
807
808
809
810
811
812
 
# File 'lib/bio/io/flatfile/index.rb', line 806

def add_record(str)
  n = records
  rs = record_size
  @file.seek(0, IO::SEEK_END)
  write_record(str)
  @records += 1
end

#close ⇒ Object 



755
756
757
758
759
760
761
762
 
# File 'lib/bio/io/flatfile/index.rb', line 755

def close
  if @file then
    DEBUG.print "FlatMappingFile: close #{@filename}\n"
    @file.close
    @file = nil
  end
  nil
end

#each ⇒ Object

export/import/edit data



841
842
843
844
845
846
847
848
 
# File 'lib/bio/io/flatfile/index.rb', line 841

def each
  n = records
  seek(0)
  (0...n).each do |i|
    yield Record.new(get_record(i))
  end
  self
end

#export_tsv(stream) ⇒ Object 



850
851
852
853
854
855
 
# File 'lib/bio/io/flatfile/index.rb', line 850

def export_tsv(stream)
  self.each do |x|
    stream << "#{x.to_s}\n"
  end
  stream
end

#get_record(i) ⇒ Object 



776
777
778
779
780
781
782
 
# File 'lib/bio/io/flatfile/index.rb', line 776

def get_record(i)
  rs = record_size
  seek(i)
  str = @file.read(rs)
  #DEBUG.print "get_record(#{i})=#{str.inspect}\n"
  str
end

#import_tsv_files(flag_primary, mode, sort_proc, *files) ⇒ Object 



948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
 
# File 'lib/bio/io/flatfile/index.rb', line 948

def import_tsv_files(flag_primary, mode, sort_proc, *files)
  require 'tempfile'

  tmpfile1 = Tempfile.open('flat')
  self.export_tsv(tmpfile1) unless mode == :new
  tmpfile1.close(false)

  tmpfile0 = Tempfile.open('sorted')
  tmpfile0.close(false)

  sort_proc.call(tmpfile0.path, tmpfile1.path, *files)

  tmpmap = self.class.new(self.filename + ".#{$$}.tmp~", 'wb+')
  tmpmap.init_with_sorted_tsv_file(tmpfile0.path, flag_primary)
  tmpmap.close
  self.close

  begin
    File.rename(self.filename, self.filename + ".#{$$}.bak~")
  rescue Errno::ENOENT
  end
  File.rename(tmpmap.filename, self.filename)
  begin
    File.delete(self.filename + ".#{$$}.bak~")
  rescue Errno::ENOENT
  end

  tmpfile0.close(true)
  tmpfile1.close(true)
  self
end

#init(rs) ⇒ Object 



827
828
829
830
831
832
833
834
835
836
837
838
 
# File 'lib/bio/io/flatfile/index.rb', line 827

def init(rs)
  unless 0 < rs and rs < 10 ** @@recsize_width then
    raise 'record size out of range'
  end
  open
  @record_size = rs
  str = sprintf("%0*d", @@recsize_width, rs)
  @file.truncate(0)
  @file.seek(0, IO::SEEK_SET)
  @file.write(str)
  @records = 0
end

#init_with_sorted_tsv_file(filename, flag_primary = false) ⇒ Object 



857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
 
# File 'lib/bio/io/flatfile/index.rb', line 857

def init_with_sorted_tsv_file(filename, flag_primary = false)
  rec_size = 1
  f = File.open(filename)
  f.each do |y|
    rec_size = y.chomp.length if rec_size < y.chomp.length
  end
  self.init(rec_size)

  prev = nil
  f.rewind
  if flag_primary then
    f.each do |y|
      x = Record.new(y.chomp, rec_size)
      if prev then
        if x.key == prev.key
          DEBUG.print "Warining: overwrote unique id #{x.key.inspect}\n"
        else
          self.add_record(prev.to_s)
        end
      end
      prev = x
    end
    self.add_record(prev.to_s) if prev
  else
    f.each do |y|
      x = Record.new(y.chomp, rec_size)
      self.add_record(x.to_s) if x != prev
      prev = x
    end
  end
  f.close
  self
end

#open ⇒ Object 



745
746
747
748
749
750
751
752
753
 
# File 'lib/bio/io/flatfile/index.rb', line 745

def open
  unless @file then
    DEBUG.print "FlatMappingFile: open #{@filename}\n"
    @file = File.open(@filename, @mode)
    true
  else
    nil
  end
end

#put_record(i, str) ⇒ Object 



814
815
816
817
818
819
820
821
822
823
824
825
 
# File 'lib/bio/io/flatfile/index.rb', line 814

def put_record(i, str)
  n = records
  rs = record_size
  if i >= n then
    @file.seek(0, IO::SEEK_END)
    @file.write(sprintf("%-*s", rs, '') * (i - n))
    @records = i + 1
  else
    seek(i)
  end
  write_record(str)
end

#record_size ⇒ Object 



764
765
766
767
768
769
770
771
772
773
774
 
# File 'lib/bio/io/flatfile/index.rb', line 764

def record_size
  unless @record_size then
    open
    @file.seek(0, IO::SEEK_SET)
    s = @file.read(@@recsize_width)
    raise 'strange record size' unless s =~ @@recsize_regex
    @record_size = s.to_i
    DEBUG.print "FlatMappingFile: record_size: #{@record_size}\n"
  end
  @record_size
end

#records ⇒ Object Also known as: size 



789
790
791
792
793
794
795
796
 
# File 'lib/bio/io/flatfile/index.rb', line 789

def records
  unless @records then
    rs = record_size
    @records = (@file.stat.size - @@recsize_width) / rs
    DEBUG.print "FlatMappingFile: records: #{@records}\n"
  end
  @records
end

#search(key) ⇒ Object

methods for searching



982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
 
# File 'lib/bio/io/flatfile/index.rb', line 982

def search(key)
  n = records
  return [] if n <= 0
  i = n / 2
  i_prev = nil
  DEBUG.print "binary search starts...\n"
  begin
    rec = Record.new(get_record(i))
    i_prev = i
    if key < rec.key then
      n = i
      i = i / 2
    elsif key > rec.key then
      i = (i + n) / 2
    else # key == rec.key
      result = [ rec.val ]
      j = i - 1
      while j >= 0 and
          (rec = Record.new(get_record(j))).key == key
        result << rec.val
        j = j - 1
      end
      result.reverse!
      j = i + 1
      while j < n and
          (rec = Record.new(get_record(j))).key == key
        result << rec.val
        j = j + 1
      end
      DEBUG.print "#{result.size} hits found!!\n"
      return result
    end
  end until i_prev == i
  DEBUG.print "no hits found\n"
  #nil
  []
end

#seek(i) ⇒ Object 



784
785
786
787
 
# File 'lib/bio/io/flatfile/index.rb', line 784

def seek(i)
  rs = record_size
  @file.seek(@@recsize_width + rs * i)
end

#write_record(str) ⇒ Object

methods for writing file



800
801
802
803
804
 
# File 'lib/bio/io/flatfile/index.rb', line 800

def write_record(str)
  rs = record_size
  rec = sprintf("%-*s", rs, str)[0..rs]
  @file.write(rec)
end