Class: Archivededup::Db

Inherits:
Object
  • Object
show all
Defined in:
lib/archivededup/db.rb

Instance Method Summary collapse

Constructor Details

#initialize(dbfile) ⇒ Db

Returns a new instance of Db.



9
10
11
 
# File 'lib/archivededup/db.rb', line 9

def initialize(dbfile)
  @db = DBM.new(dbfile, 0755, DBM::WRCREAT)
end

Instance Method Details

#add_directory(dir, &blk) ⇒ Object 



35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
 
# File 'lib/archivededup/db.rb', line 35

def add_directory(dir, &blk)
  Dir.new(dir).each do |dent|
    d = File.join(dir, dent)

    next if dent == '.'
    next if dent == '..'
    next if dent.start_with? '.'

    if File.directory? d
      add_directory(d, &blk)
    else
      yield d if block_given?
    end
  end
end

#add_file(d) ⇒ Object 



51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
 
# File 'lib/archivededup/db.rb', line 51

def add_file(d)
  # puts "Got #{d}"
  s = File.stat(d)
  # noimplemented - s.birthtime
  s.mtime
  s.ctime
  # unused - s.atime
  # puts d, s
  hash = Digest::MD5.new
  outbuf = ''

  File.open(d) do |io|
    while io.read(4096, outbuf) do
      hash.update outbuf
    end
  end
  
  unless @db.has_key? "name - #{d}"

    hash = hash.hexdigest

    @db["name - #{d}"] = YAML::dump({
      dir: d,
      hash: hash,
    })

    if @db.has_key?('hash - '+hash)
      o = YAML::load(@db['hash - '+hash])
      buf1 = ''
      buf2 = ''
      File.open(d) do |io1|
        File.open(o['files'][0]) do |io2|
          v1 = ''
          v2 = ''
          while v1 && v2 && v1 == v2 do
            v1 = io1.read(4096, buf1)
            v2 = io2.read(4096, buf2)
          end

          if v1 == v2
            o['files'] << d
            puts "Dups found: #{o}"
            @db['hash - '+hash] = YAML::dump(o)
          end
        end
      end
    else
      puts "Storing #{d}."
      @db['hash - '+hash] = YAML::dump({
        'files' => [ d ],
        'hash' => hash,
      })
    end
  else
    puts "Had #{d}."
  end
end

#close ⇒ Object 



109
110
111
 
# File 'lib/archivededup/db.rb', line 109

def close()
  @db.close()
end

#each_dup ⇒ Object 



23
24
25
26
27
28
29
30
31
32
33
 
# File 'lib/archivededup/db.rb', line 23

def each_dup
  @db.select do |k, v|
    k.start_with? 'hash - '
  end.map do |k, v|
    [k, YAML::load(v)]
  end.select do |k, r|
    r['files'].length > 1
  end.each do |k, r|
    yield k, r
  end
end

#each_hash ⇒ Object 



13
14
15
16
17
18
19
20
21
 
# File 'lib/archivededup/db.rb', line 13

def each_hash
  @db.select do |k, v|
    k.start_with? 'hash - '
  end.map do |k, v|
    [k, YAML::load(v)]
  end.each do |k, r|
    yield k, r
  end
end