File: fs_bucket.rb

package info (click to toggle)
ruby-mongo 2.5.1-2
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid, trixie
  • size: 4,332 kB
  • sloc: ruby: 45,579; makefile: 5
file content (455 lines) | stat: -rw-r--r-- 16,085 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
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
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
123
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
166
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
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
# Copyright (C) 2014-2017 MongoDB, Inc.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#   http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

module Mongo
  module Grid

    # Represents a view of the GridFS in the database.
    #
    # @since 2.0.0
    class FSBucket
      extend Forwardable

      # The default root prefix.
      #
      # @since 2.0.0
      DEFAULT_ROOT = 'fs'.freeze

      # The specification for the chunks collection index.
      #
      # @since 2.0.0
      CHUNKS_INDEX = { :files_id => 1, :n => 1 }.freeze

      # The specification for the files collection index.
      #
      # @since 2.1.0
      FILES_INDEX = { filename: 1, uploadDate: 1 }.freeze

      # @return [ Collection ] chunks_collection The chunks collection.
      #
      # @since 2.0.0
      attr_reader :chunks_collection

      # @return [ Database ] database The database.
      #
      # @since 2.0.0
      attr_reader :database

      # @return [ Collection ] files_collection The files collection.
      #
      # @since 2.0.0
      attr_reader :files_collection

      # @return [ Hash ] options The FSBucket options.
      #
      # @since 2.1.0
      attr_reader :options

      # Get client from the database.
      #
      # @since 2.1.0
      def_delegators :database,
                     :client

      # Find files collection documents matching a given selector.
      #
      # @example Find files collection documents by a filename.
      #   fs.find(filename: 'file.txt')
      #
      # @param [ Hash ] selector The selector to use in the find.
      # @param [ Hash ] options The options for the find.
      #
      # @option options [ Integer ] :batch_size The number of documents returned in each batch
      #   of results from MongoDB.
      # @option options [ Integer ] :limit The max number of docs to return from the query.
      # @option options [ true, false ] :no_cursor_timeout The server normally times out idle
      #   cursors after an inactivity period (10 minutes) to prevent excess memory use.
      #   Set this option to prevent that.
      # @option options [ Integer ] :skip The number of docs to skip before returning results.
      # @option options [ Hash ] :sort The key and direction pairs by which the result set
      #   will be sorted.
      #
      # @return [ CollectionView ] The collection view.
      #
      # @since 2.1.0
      def find(selector = nil, options = {})
        opts = options.merge(read: read_preference) if read_preference
        files_collection.find(selector, opts || options)
      end

      # Find a file in the GridFS.
      #
      # @example Find a file by its id.
      #   fs.find_one(_id: id)
      #
      # @example Find a file by its filename.
      #   fs.find_one(filename: 'test.txt')
      #
      # @param [ Hash ] selector The selector.
      #
      # @return [ Grid::File ] The file.
      #
      # @since 2.0.0
      #
      # @deprecated Please use #find instead with a limit of -1.
      #   Will be removed in version 3.0.
      def find_one(selector = nil)
        file_info = files_collection.find(selector).first
        return nil unless file_info
        chunks = chunks_collection.find(:files_id => file_info[:_id]).sort(:n => 1)
        Grid::File.new(chunks.to_a, Options::Mapper.transform(file_info, Grid::File::Info::MAPPINGS.invert))
      end

      # Insert a single file into the GridFS.
      #
      # @example Insert a single file.
      #   fs.insert_one(file)
      #
      # @param [ Grid::File ] file The file to insert.
      #
      # @return [ BSON::ObjectId ] The file id.
      #
      # @since 2.0.0
      #
      # @deprecated Please use #upload_from_stream or #open_upload_stream instead.
      #   Will be removed in version 3.0.
      def insert_one(file)
        @indexes ||= ensure_indexes!
        chunks_collection.insert_many(file.chunks)
        files_collection.insert_one(file.info)
        file.id
      end

      # Create the GridFS.
      #
      # @example Create the GridFS.
      #   Grid::FSBucket.new(database)
      #
      # @param [ Database ] database The database the files reside in.
      # @param [ Hash ] options The GridFS options.
      #
      # @option options [ String ] :fs_name The prefix for the files and chunks
      #   collections.
      # @option options [ String ] :bucket_name The prefix for the files and chunks
      #   collections.
      # @option options [ Integer ] :chunk_size Override the default chunk
      #   size.
      # @option options [ String ] :write The write concern.
      # @option options [ String ] :read The read preference.
      #
      # @since 2.0.0
      def initialize(database, options = {})
        @database = database
        @options = options
        @chunks_collection = database[chunks_name]
        @files_collection = database[files_name]
      end

      # Get the prefix for the GridFS
      #
      # @example Get the prefix.
      #   fs.prefix
      #
      # @return [ String ] The GridFS prefix.
      #
      # @since 2.0.0
      def prefix
        @options[:fs_name] || @options[:bucket_name]|| DEFAULT_ROOT
      end

      # Remove a single file from the GridFS.
      #
      # @example Remove a file from the GridFS.
      #   fs.delete_one(file)
      #
      # @param [ Grid::File ] file The file to remove.
      #
      # @return [ Result ] The result of the remove.
      #
      # @since 2.0.0
      def delete_one(file)
        delete(file.id)
      end

      # Remove a single file, identified by its id from the GridFS.
      #
      # @example Remove a file from the GridFS.
      #   fs.delete(id)
      #
      # @param [ BSON::ObjectId, Object ] id The id of the file to remove.
      #
      # @return [ Result ] The result of the remove.
      #
      # @raise [ Error::FileNotFound ] If the file is not found.
      #
      # @since 2.1.0
      def delete(id)
        result = files_collection.find({ :_id => id }, @options).delete_one
        chunks_collection.find({ :files_id => id }, @options).delete_many
        raise Error::FileNotFound.new(id, :id) if result.n == 0
        result
      end

      # Opens a stream from which a file can be downloaded, specified by id.
      #
      # @example Open a stream from which a file can be downloaded.
      #   fs.open_download_stream(id)
      #
      # @param [ BSON::ObjectId, Object ] id The id of the file to read.
      #
      # @return [ Stream::Read ] The stream to read from.
      #
      # @yieldparam [ Hash ] The read stream.
      #
      # @since 2.1.0
      def open_download_stream(id)
        read_stream(id).tap do |stream|
          if block_given?
            yield stream
            stream.close
          end
        end
      end

      # Downloads the contents of the file specified by id and writes them to
      # the destination io object.
      #
      # @example Download the file and write it to the io object.
      #   fs.download_to_stream(id, io)
      #
      # @param [ BSON::ObjectId, Object ] id The id of the file to read.
      # @param [ IO ] io The io object to write to.
      #
      # @since 2.1.0
      def download_to_stream(id, io)
        open_download_stream(id) do |stream|
          stream.each do |chunk|
            io << chunk
          end
        end
      end

      # Opens a stream from which the application can read the contents of the stored file
      # specified by filename and the revision in options.
      #
      # Revision numbers are defined as follows:
      # 0 = the original stored file
      # 1 = the first revision
      # 2 = the second revision
      # etc…
      # -2 = the second most recent revision
      # -1 = the most recent revision
      #
      # @example Open a stream to download the most recent revision.
      #   fs.open_download_stream_by_name('some-file.txt')
      #
      # # @example Open a stream to download the original file.
      #   fs.open_download_stream_by_name('some-file.txt', revision: 0)
      #
      # @example Open a stream to download the second revision of the stored file.
      #   fs.open_download_stream_by_name('some-file.txt', revision: 2)
      #
      # @param [ String ] filename The file's name.
      # @param [ Hash ] opts Options for the download.
      #
      # @option opts [ Integer ] :revision The revision number of the file to download.
      #   Defaults to -1, the most recent version.
      #
      # @return [ Stream::Read ] The stream to read from.
      #
      # @raise [ Error::FileNotFound ] If the file is not found.
      # @raise [ Error::InvalidFileRevision ] If the requested revision is not found for the file.
      #
      # @yieldparam [ Hash ] The read stream.
      #
      # @since 2.1.0
      def open_download_stream_by_name(filename, opts = {}, &block)
        revision = opts.fetch(:revision, -1)
        if revision < 0
          skip = revision.abs - 1
          sort = { 'uploadDate' => Mongo::Index::DESCENDING }
        else
          skip = revision
          sort = { 'uploadDate' => Mongo::Index::ASCENDING }
        end
        file_doc = files_collection.find({ filename: filename} ,
                                           projection: { _id: 1 },
                                           sort: sort,
                                           skip: skip,
                                           limit: -1).first
        unless file_doc
          raise Error::FileNotFound.new(filename, :filename) unless opts[:revision]
          raise Error::InvalidFileRevision.new(filename, opts[:revision])
        end
        open_download_stream(file_doc[:_id], &block)
      end

      # Downloads the contents of the stored file specified by filename and by the
      # revision in options and writes the contents to the destination io object.
      #
      # Revision numbers are defined as follows:
      # 0 = the original stored file
      # 1 = the first revision
      # 2 = the second revision
      # etc…
      # -2 = the second most recent revision
      # -1 = the most recent revision
      #
      # @example Download the most recent revision.
      #   fs.download_to_stream_by_name('some-file.txt', io)
      #
      # # @example Download the original file.
      #   fs.download_to_stream_by_name('some-file.txt', io, revision: 0)
      #
      # @example Download the second revision of the stored file.
      #   fs.download_to_stream_by_name('some-file.txt', io, revision: 2)
      #
      # @param [ String ] filename The file's name.
      # @param [ IO ] io The io object to write to.
      # @param [ Hash ] opts Options for the download.
      #
      # @option opts [ Integer ] :revision The revision number of the file to download.
      #   Defaults to -1, the most recent version.
      #
      # @raise [ Error::FileNotFound ] If the file is not found.
      # @raise [ Error::InvalidFileRevision ] If the requested revision is not found for the file.
      #
      # @since 2.1.0
      def download_to_stream_by_name(filename, io, opts = {})
        download_to_stream(open_download_stream_by_name(filename, opts).file_id, io)
      end

      # Opens an upload stream to GridFS to which the contents of a user file came be written.
      #
      # @example Open a stream to which the contents of a file came be written.
      #   fs.open_upload_stream('a-file.txt')
      #
      # @param [ String ] filename The filename of the file to upload.
      # @param [ Hash ] opts The options for the write stream.
      #
      # @option opts [ Object ] :file_id An optional unique file id. An ObjectId is generated otherwise.
      # @option opts [ Integer ] :chunk_size Override the default chunk size.
      # @option opts [ Hash ] :write The write concern.
      # @option opts [ Hash ] :metadata User data for the 'metadata' field of the files
      #   collection document.
      # @option opts [ String ] :content_type The content type of the file.
      #   Deprecated, please use the metadata document instead.
      # @option opts [ Array<String> ] :aliases A list of aliases.
      #   Deprecated, please use the metadata document instead.
      #
      # @return [ Stream::Write ] The write stream.
      #
      # @yieldparam [ Hash ] The write stream.
      #
      # @since 2.1.0
      def open_upload_stream(filename, opts = {})
        write_stream(filename, opts).tap do |stream|
          if block_given?
            yield stream
            stream.close
          end
        end
      end

      # Uploads a user file to a GridFS bucket.
      # Reads the contents of the user file from the source stream and uploads it as chunks in the
      # chunks collection. After all the chunks have been uploaded, it creates a files collection
      # document for the filename in the files collection.
      #
      # @example Upload a file to the GridFS bucket.
      #   fs.upload_from_stream('a-file.txt', file)
      #
      # @param [ String ] filename The filename of the file to upload.
      # @param [ IO ] io The source io stream to upload from.
      # @param [ Hash ] opts The options for the write stream.
      #
      # @option opts [ Object ] :file_id An optional unique file id. An ObjectId is generated otherwise.
      # @option opts [ Integer ] :chunk_size Override the default chunk size.
      # @option opts [ Hash ] :write The write concern.
      # @option opts [ Hash ] :metadata User data for the 'metadata' field of the files
      #   collection document.
      # @option opts [ String ] :content_type The content type of the file. Deprecated, please
      #   use the metadata document instead.
      # @option opts [ Array<String> ] :aliases A list of aliases. Deprecated, please use the
      #   metadata document instead.
      #
      # @return [ BSON::ObjectId ] The ObjectId file id.
      #
      # @since 2.1.0
      def upload_from_stream(filename, io, opts = {})
        open_upload_stream(filename, opts) do |stream|
          begin
            stream.write(io)
          rescue IOError
            begin
              stream.abort
            rescue Error::OperationFailure
            end
            raise
          end
        end.file_id
      end

      # Get the read preference.
      #
      # @example Get the read preference.
      #   fs.read_preference
      #
      # @return [ Mongo::ServerSelector ] The read preference.
      #
      # @since 2.1.0
      def read_preference
        @read_preference ||= options[:read] || database.read_preference
      end

      # Get the write concern.
      #
      # @example Get the write concern.
      #   stream.write_concern
      #
      # @return [ Mongo::WriteConcern ] The write concern.
      #
      # @since 2.1.0
      def write_concern
        @write_concern ||= @options[:write] ? WriteConcern.get(@options[:write]) :
            database.write_concern
      end

      private

      def read_stream(id)
        Stream.get(self, Stream::READ_MODE, { file_id: id }.merge!(options))
      end

      def write_stream(filename, opts)
        Stream.get(self, Stream::WRITE_MODE, { filename: filename }.merge!(options).merge!(opts))
      end

      def chunks_name
        "#{prefix}.#{Grid::File::Chunk::COLLECTION}"
      end

      def files_name
        "#{prefix}.#{Grid::File::Info::COLLECTION}"
      end

      def ensure_indexes!
        if files_collection.find({}, limit: 1, projection: { _id: 1 }).first.nil?
          chunks_collection.indexes.create_one(FSBucket::CHUNKS_INDEX, :unique => true)
          files_collection.indexes.create_one(FSBucket::FILES_INDEX)
        end
      end
    end
  end
end