Module: Persist

Defined in:
lib/rbbt/persist.rb,
lib/rbbt/persist/tsv.rb,
lib/rbbt/persist/tsv/cdb.rb,
lib/rbbt/persist/tsv/lmdb.rb,
lib/rbbt/persist/tsv/leveldb.rb,
lib/rbbt/persist/tsv/kyotocabinet.rb,
lib/rbbt/persist/tsv/tokyocabinet.rb

Defined Under Namespace

Modules: CDBAdapter, KCAdapter, LMDBAdapter, LevelDBAdapter, TCAdapter

Constant Summary collapse

MEMORY = 
{}
MAX_FILE_LENGTH = 
150
TRUE_STRINGS = 
Set.new ["true", "True", "TRUE", "t", "T", "1", "yes", "Yes", "YES", "y", "Y", "ON", "on"]
CONNECTIONS = 
{}

Class Attribute Summary collapse

Class Method Summary collapse

Class Attribute Details

.cachedir ⇒ Object

Returns the value of attribute cachedir.



10
11
12
 
# File 'lib/rbbt/persist.rb', line 10

def cachedir
  @cachedir
end

.lock_dir ⇒ Object

Returns the value of attribute lock_dir.



18
19
20
 
# File 'lib/rbbt/persist.rb', line 18

def lock_dir
  @lock_dir
end

Class Method Details

.get_filename(source) ⇒ Object 



34
35
36
37
38
39
40
41
42
43
44
45
 
# File 'lib/rbbt/persist/tsv.rb', line 34

def self.get_filename(source)
  case
  when Path === source
    source
  when (source.respond_to?(:filename) and source.filename)
    source.filename
  when source.respond_to?(:cmd)
    "CMD-#{Misc.digest(source.cmd)}"
  when TSV === source
    "TSV[#{Misc.digest Misc.fingerprint(source)}]"
  end || source.object_id.to_s
end

.is_persisted?(path, persist_options = {}) ⇒ Boolean

Returns:

  • (Boolean) 


34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
 
# File 'lib/rbbt/persist.rb', line 34

def self.is_persisted?(path, persist_options = {})
  return false if not Open.exists? path
  return false if TrueClass === persist_options[:update]

  check = persist_options[:check]
  if not check.nil?
    if Array === check
      return false if check.select{|file| newer? path, file}.any?
   else
      return false if newer? path, check
   end
  end

  return true
end

.load_file(path, type) ⇒ Object 



85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
 
# File 'lib/rbbt/persist.rb', line 85

def self.load_file(path, type)
  case (type || "nil").to_sym
  when :nil
    nil
  when :boolean
    TRUE_STRINGS.include? Open.read(path).chomp.strip
  when :annotations
    Annotated.load_tsv TSV.open(path)
  when :tsv
    TSV.open(path)
  when :marshal_tsv
    TSV.setup(Marshal.load(Open.open(path)))
  when :fwt
    FixWidthTable.get(path) 
  when :string, :text
    Open.read(path)
  when :binary
    f = File.open(path, 'rb')
    res = f.read
    f.close
    res.force_encoding("ASCII-8BIT") if res.respond_to? :force_encoding
    res
  when :array
    res = Open.read(path).split("\n", -1)
    res.pop
    res
  when :marshal
    Marshal.load(Open.open(path))
  when :yaml
    YAML.load(Open.open(path))
  when :float
    Open.read(path).to_f
  when :integer
    Open.read(path).to_i
  else
    raise "Unknown persistence: #{ type }"
  end
end

.memory(name, options = {}, &block) ⇒ Object 



303
304
305
306
307
308
309
 
# File 'lib/rbbt/persist.rb', line 303

def self.memory(name, options = {}, &block)
  file = name
  file << "_" << Misc.hash2md5(options) if options.any?
  options = Misc.add_defaults options, :persist => true, :file => file

  persist name, :memory, options, &block
end

.newer?(path, file) ⇒ Boolean

Returns:

  • (Boolean) 


28
29
30
31
32
 
# File 'lib/rbbt/persist.rb', line 28

def self.newer?(path, file)
  return true if not Open.exists? file
  return true if File.mtime(path) < File.mtime(file)
  return false
end

.open_cdb(path, write, serializer = nil) ⇒ Object 



125
126
127
128
129
130
131
132
133
134
135
136
137
138
 
# File 'lib/rbbt/persist/tsv/cdb.rb', line 125

def self.open_cdb(path, write, serializer = nil)
  write = true unless File.exists? path

  FileUtils.mkdir_p File.dirname(path) unless File.exists?(File.dirname(path))

  database = Persist::CDBAdapter.open(path, write)

  unless serializer == :clean
    TSV.setup database
    database.serializer = serializer || database.serializer
  end

  database
end

.open_database(path, write, serializer = nil, type = "HDB") ⇒ Object 



47
48
49
50
51
52
53
54
55
56
57
58
59
60
 
# File 'lib/rbbt/persist/tsv.rb', line 47

def self.open_database(path, write, serializer = nil, type = "HDB")
  case type
  when "LevelDB"
    Persist.open_leveldb(path, write, serializer)
  when "CDB"
    Persist.open_cdb(path, write, serializer)
  when "LMDB"
    Persist.open_lmdb(path, write, serializer)
  when 'kch', 'kct'
    Persist.open_kyotocabinet(path, write, serializer, type)
  else
    Persist.open_tokyocabinet(path, write, serializer, type)
  end
end

.open_kyotocabinet(path, write, serializer = nil, kyotocabinet_class = 'kch') ⇒ Object 



154
155
156
157
158
159
160
161
162
163
164
165
166
167
 
# File 'lib/rbbt/persist/tsv/kyotocabinet.rb', line 154

def self.open_kyotocabinet(path, write, serializer = nil,  kyotocabinet_class= 'kch')
  write = true unless File.exists? path

  FileUtils.mkdir_p File.dirname(path) unless File.exists?(File.dirname(path))

  database = Persist::KCAdapter.open(path, write, kyotocabinet_class)

  unless serializer == :clean
    TSV.setup database
    database.serializer = serializer || database.serializer
  end

  database
end

.open_leveldb(path, write, serializer = nil) ⇒ Object 



106
107
108
109
110
111
112
113
114
115
116
117
118
119
 
# File 'lib/rbbt/persist/tsv/leveldb.rb', line 106

def self.open_leveldb(path, write, serializer = nil)
  write = true unless File.exists? path

  FileUtils.mkdir_p File.dirname(path) unless File.exists?(File.dirname(path))

  database = Persist::LevelDBAdapter.open(path, write)

  unless serializer == :clean
    TSV.setup database
    database.serializer = serializer || database.serializer
  end

  database
end

.open_lmdb(path, write, serializer = nil) ⇒ Object 



134
135
136
137
138
139
140
141
142
143
144
145
146
147
 
# File 'lib/rbbt/persist/tsv/lmdb.rb', line 134

def self.open_lmdb(path, write, serializer = nil)
  write = true unless File.exists? path

  FileUtils.mkdir_p File.dirname(path) unless File.exists?(File.dirname(path))

  database = Persist::LMDBAdapter.open(path, write)

  unless serializer == :clean
    TSV.setup database
    database.serializer = serializer || database.serializer
  end

  database
end

.open_tokyocabinet(path, write, serializer = nil, tokyocabinet_class = TokyoCabinet::HDB) ⇒ Object 



151
152
153
154
155
156
157
158
159
160
161
162
163
164
 
# File 'lib/rbbt/persist/tsv/tokyocabinet.rb', line 151

def self.open_tokyocabinet(path, write, serializer = nil, tokyocabinet_class = TokyoCabinet::HDB)
  write = true unless File.exists? path

  FileUtils.mkdir_p File.dirname(path) unless File.exists?(File.dirname(path))

  database = Persist::TCAdapter.open(path, write, tokyocabinet_class)

  unless serializer == :clean
    TSV.setup database
    database.serializer = serializer || database.serializer
  end

  database
end

.persist(name, type = nil, persist_options = {}) ⇒ Object 



167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
 
# File 'lib/rbbt/persist.rb', line 167

def self.persist(name, type = nil, persist_options = {})
  type ||= :marshal
  persist_options = Misc.add_defaults persist_options, :persist => true
  other_options = Misc.process_options persist_options, :other

  if persist_options[:persist]
    path = persistence_path(name, persist_options, other_options || {})

    case 
    when type.to_sym === :memory
      Persist::MEMORY[path] ||= yield

    when (type.to_sym == :annotations and persist_options.include? :annotation_repo)

      repo = persist_options[:annotation_repo]

      keys = nil
      subkey = name + ":"

      if String === repo
        repo = Persist.open_tokyocabinet(repo, false, :list, "BDB")
        repo.read_and_close do
          keys = repo.range subkey + 0.chr, true, subkey + 254.chr, true
        end
        repo.close
      else
        repo.read_and_close do
          keys = repo.range subkey + 0.chr, true, subkey + 254.chr, true
        end
      end

      case
      when (keys.length == 1 and keys.first == subkey + 'NIL')
        nil
      when (keys.length == 1 and keys.first == subkey + 'EMPTY')
        []
      when (keys.length == 1 and keys.first =~ /:SINGLE$/)
        key = keys.first
        values = repo.read_and_close do
          repo[key]
        end
        Annotated.load_tsv_values(key, values, "literal", "annotation_types", "JSON")
      when (keys.any? and not keys.first =~ /ANNOTATED_DOUBLE_ARRAY/)
        repo.read_and_close do
          keys.sort_by{|k| k.split(":").last.to_i}.collect{|key|
            v = repo[key]
            Annotated.load_tsv_values(key, v, "literal", "annotation_types", "JSON")
          }
        end
      when (keys.any? and keys.first =~ /ANNOTATED_DOUBLE_ARRAY/)
        repo.read_and_close do

          res = keys.sort_by{|k| k.split(":").last.to_i}.collect{|key|
            v = repo[key]
            Annotated.load_tsv_values(key, v, "literal", "annotation_types", "JSON")
          }

          res.first.annotate res
          res.extend AnnotatedArray

          res
        end
      else
        entities = yield

        repo.write_and_close do 
          case
          when entities.nil?
            repo[subkey + "NIL"] = nil
          when entities.empty?
            repo[subkey + "EMPTY"] = nil
          when (not Array === entities or (AnnotatedArray === entities and not Array === entities.first))
            tsv_values = entities.tsv_values("literal", "annotation_types", "JSON") 
            repo[subkey + entities.id << ":" << "SINGLE"] = tsv_values
          when (not Array === entities or (AnnotatedArray === entities and AnnotatedArray === entities.first))
            entities.each_with_index do |e,i|
              next if e.nil?
              tsv_values = e.tsv_values("literal", "annotation_types", "JSON") 
              repo[subkey + e.id << ":ANNOTATED_DOUBLE_ARRAY:" << i.to_s] = tsv_values
            end
          else
            entities.each_with_index do |e,i|
              next if e.nil?
              tsv_values = e.tsv_values("literal", "annotation_types", "JSON") 
              repo[subkey + e.id << ":" << i.to_s] = tsv_values
            end
          end
        end

        entities
      end

    else

      if is_persisted?(path, persist_options)
        Log.low "Persist up-to-date: #{ path } - #{Misc.fingerprint persist_options}"
        return path if persist_options[:no_load]
        return load_file(path, type) 
      end

      begin
        lock_filename = Persist.persistence_path(path + '.persist', {:dir => Persist.lock_dir})
        Misc.lock lock_filename  do
          if is_persisted?(path, persist_options)
            Log.low "Persist up-to-date (suddenly): #{ path } - #{Misc.fingerprint persist_options}"
            return path if persist_options[:no_load]
            return load_file(path, type) 
          end

          Log.medium "Persist create: #{ path } - #{persist_options.inspect[0..100]}"
          res = yield

          if res.nil?
            res = load_file(path) unless persist_options[:no_load]
          else
            Misc.lock(path) do
              save_file(path, type, res)
            end
          end

          return path if persist_options[:no_load]

          res
        end
      rescue
        Log.high "Error in persist: #{path}#{Open.exists?(path) ? Log.color(:red, " Erasing") : ""}"
        FileUtils.rm path if Open.exists? path 
        raise $!
      end
    end

  else
    yield
  end
end

.persist_tsv(source, filename, options = {}, persist_options = {}, &block) ⇒ Object 



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
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
 
# File 'lib/rbbt/persist/tsv.rb', line 62

def self.persist_tsv(source, filename, options = {}, persist_options = {}, &block)
  persist_options[:prefix] ||= "TSV"

  if data = persist_options[:data]
    yield data
    return data 
  end

  filename ||= get_filename(source)

  if not persist_options[:persist]
    data = {}

    yield(data) 

    return data 
  end

  path = persistence_path(filename, persist_options, options)

  lock_filename = Persist.persistence_path(path, {:dir => TSV.lock_dir})

  if is_persisted?(path) and not persist_options[:update]
    Log.debug "TSV persistence up-to-date: #{ path }"
    return open_database(path, false, nil, persist_options[:engine] || TokyoCabinet::HDB) 
  end

  Misc.lock lock_filename do
    begin
      if is_persisted?(path) and not persist_options[:update]
        Log.debug "TSV persistence (suddenly) up-to-date: #{ path }"
        return open_database(path, false, nil, persist_options[:engine] || TokyoCabinet::HDB) 
      end

      FileUtils.rm path if File.exists? path

      Log.medium "TSV persistence creating: #{ path }"

      tmp_path = path + '.persist'

      data = open_database(tmp_path, true, persist_options[:serializer], persist_options[:engine] || TokyoCabinet::HDB)
      data.serializer = :type if TSV === data and data.serializer.nil?

      data.write_and_read do
        yield data
      end

      FileUtils.mv data.persistence_path, path if File.exists? data.persistence_path and not File.exists? path
      tsv = CONNECTIONS[path] = CONNECTIONS.delete tmp_path
      tsv.persistence_path = path
      tsv.fix_io if tsv.respond_to? :fix_io

      data
    rescue Exception
      Log.error "Captured error during persist_tsv. Erasing: #{path}"
      FileUtils.rm_rf tmp_path if tmp_path and File.exists? tmp_path
      FileUtils.rm_rf path if path and File.exists? path
      raise $!
    end
  end
end

.persistence_path(file, persist_options = {}, options = {}) ⇒ Object 



50
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
 
# File 'lib/rbbt/persist.rb', line 50

def self.persistence_path(file, persist_options = {}, options = {})
  persistence_file = Misc.process_options persist_options, :file
  return persistence_file unless persistence_file.nil?

  prefix = Misc.process_options persist_options, :prefix

  if prefix.nil?
    perfile = file.to_s.gsub(/\//, '>') 
  else
    perfile = prefix.to_s + ":" + file.to_s.gsub(/\//, '>') 
  end

  if options.include? :filters
    options[:filters].each do |match,value|
      perfile = perfile + "&F[#{match}=#{Misc.digest(value.inspect)}]"
    end
  end

  persistence_dir = Misc.process_options(persist_options, :dir) || Persist.cachedir 
  Path.setup(persistence_dir) unless Path === persistence_dir

  filename = perfile.gsub(/\s/,'_').gsub(/\//,'>')
  clean_options = options
  clean_options.delete :unnamed
  clean_options.delete "unnamed"

  filename = filename[0..MAX_FILE_LENGTH] << Misc.digest(filename[MAX_FILE_LENGTH+1..-1]) if filename.length > MAX_FILE_LENGTH + 10

  options_md5 = Misc.hash2md5 clean_options
  filename  << ":" << options_md5 unless options_md5.empty?

  persistence_dir[filename].find
end

.save_file(path, type, content) ⇒ Object 



124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
 
# File 'lib/rbbt/persist.rb', line 124

def self.save_file(path, type, content)

  return if content.nil?
  
  case (type || "nil").to_sym
  when :nil
    nil
  when :boolean
    Misc.sensiblewrite(path, content ? "true" : "false")
  when :fwt
    content.file.seek 0
    Misc.sensiblewrite(path, content.file.read)
  when :tsv
    Misc.sensiblewrite(path, content.to_s)
  when :annotations
    Misc.sensiblewrite(path, Annotated.tsv(content, :all).to_s)
  when :string, :text
    Misc.sensiblewrite(path, content)
  when :binary
    content.force_encoding("ASCII-8BIT") if content.respond_to? :force_encoding
    f = File.open(path, 'wb')
    f.puts content
    f.close
    content
  when :array
    if content.empty?
      Misc.sensiblewrite(path, "")
    else
      Misc.sensiblewrite(path, content * "\n" + "\n")
    end
  when :marshal_tsv
    Misc.sensiblewrite(path, Marshal.dump(content.dup))
  when :marshal
    Misc.sensiblewrite(path, Marshal.dump(content))
  when :yaml
    Misc.sensiblewrite(path, YAML.dump(content))
  when :float, :integer, :tsv
    Misc.sensiblewrite(path, content.to_s)
  else
    raise "Unknown persistence: #{ type }"
  end
end