sdr-client 0.3.0 → 0.7.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 80a6d11a266d546eb1e0763b99b13b18d9f64c558ef21db50f3f509623a4f4bf
-  data.tar.gz: 5ec2344617583af3322f0426e39328aee355ee4c12d7338d8734f529110e8da6
+  metadata.gz: bb3f20e21233ef1ef0f4d72f4d7dcd0e4bb52968b955152e62259e023b730b7b
+  data.tar.gz: ecbb08752fcd3c71b4868404bc48871eaba138c852304678a466cd85de3fe4bd
 SHA512:
-  metadata.gz: 3a145ef20a1c42abaed33ab79746778844e922b530a6c48d1b7f6d6f50a0e5a2c3f23e15a804fc5b75b3baeaac6baa4e67b35391965b5bb5da34892b345c9805
-  data.tar.gz: 2600c209c4f35f2ab7a8d7209189c690ec8f06de819666fd7bdd38f63290b997aa87355d54c7a5b5b2e4633ecfd02f75e908815923c2f44116373a999789bf68
+  metadata.gz: '08307c141dd01737019f40e1a999b8fd9aa41abca57379f655a9ac5773eba55f974c4d3c34c804e44cd7c8a8792c292e84f5d8c7bf62dd2021b32dd87832420b'
+  data.tar.gz: 928ba2d1377cf11bc1e1cfaff2891748ac916d3a3d9e83fca0ac07e0efdbd450ceb895a066bcc3c3a7a4355f186717efb4f5c4d6540867336a21095064f02cb2

data/.rubocop.yml

@@ -13,3 +13,9 @@ Layout/LineLength:
 Metrics/BlockLength:
   Exclude:
     - 'spec/**/*'
+  ExcludedMethods:
+    - 'OptionParser.new'
+
+Naming/FileName:
+  Exclude:
+    - 'lib/sdr-client.rb'

data/.rubocop_todo.yml

@@ -14,8 +14,3 @@ Metrics/AbcSize:
 # Configuration parameters: CountComments, ExcludedMethods.
 Metrics/MethodLength:
   Max: 13
-
-# Offense count: 2
-# Configuration parameters: CountKeywordArgs.
-Metrics/ParameterLists:
-  Max: 8

data/README.md

@@ -8,6 +8,10 @@
 This is a CLI for interacting with the Stanford Digital Repository API.
 The code for the SDR API server is at https://github.com/sul-dlss/sdr-api
 
+## Install
+
+`gem install sdr-client`
+
 ## Usage
 
 Log in:

data/exe/sdr

@@ -3,6 +3,7 @@
 
 $LOAD_PATH.unshift 'lib'
 require 'optparse'
+require 'sdr_client'
 
 options = {}
 global = OptionParser.new do |opts|
@@ -60,6 +61,20 @@ subcommands = {
       options[:source_id] = source_id
     end
 
+    opts.on('--strategy STRATEGY',
+            'The strategy to use for distributing files into filesets. Either "default" or "filename"') do |strategy|
+      strategy_class = case strategy
+                       when 'filename'
+                         SdrClient::Deposit::MatchingFileGroupingStrategy
+                       when 'default'
+                         SdrClient::Deposit::SingleFileGroupingStrategy
+                       else
+                         warn "Unknown strategy #{strategy}"
+                         exit(1)
+                       end
+      options[:grouping_strategy] = strategy_class
+    end
+
     opts.on('-h', '--help', 'Display this screen') do
       puts opts
       exit
@@ -75,6 +90,5 @@ end
 
 subcommands[command].order!
 
-require 'sdr_client'
 options[:files] = ARGV unless ARGV.empty?
 SdrClient::CLI.start(command, options)

data/lib/sdr-client.rb

@@ -0,0 +1,3 @@
+# frozen_string_literal: true
+
+require 'sdr_client'

data/lib/sdr_client/deposit.rb

@@ -3,13 +3,17 @@
 module SdrClient
   # The namespace for the "deposit" command
   module Deposit
+    # rubocop:disable Metrics/ParameterLists
     def self.run(label: nil,
                  type: 'http://cocina.sul.stanford.edu/models/book.jsonld',
                  apo:,
                  collection:,
                  catkey: nil,
                  source_id:,
-                 url:, files: [])
+                 url:,
+                 files: [],
+                 files_metadata: {},
+                 grouping_strategy: SingleFileGroupingStrategy)
       token = Credentials.read
 
       metadata = Request.new(label: label,
@@ -18,11 +22,15 @@ module SdrClient
                              collection: collection,
                              source_id: source_id,
                              catkey: catkey)
-      Process.new(metadata: metadata, url: url, token: token, files: files).run
+      Process.new(metadata: metadata, url: url, token: token,
+                  files: files, files_metadata: files_metadata, grouping_strategy: grouping_strategy).run
     end
+    # rubocop:enable Metrics/ParameterLists
   end
 end
 require 'json'
+require 'sdr_client/deposit/single_file_grouping_strategy'
+require 'sdr_client/deposit/matching_file_grouping_strategy'
 require 'sdr_client/deposit/files/direct_upload_request'
 require 'sdr_client/deposit/files/direct_upload_response'
 require 'sdr_client/deposit/file'

data/lib/sdr_client/deposit/file.rb

@@ -4,15 +4,23 @@ module SdrClient
   module Deposit
     # This represents the File metadata that we send to the server for doing a deposit
     class File
-      def initialize(external_identifier:, label:, filename:, access: 'dark', preserve: false, shelve: false)
+      # rubocop:disable Metrics/ParameterLists
+      def initialize(external_identifier:, label:, filename:,
+                     access: 'dark', preserve: false, shelve: false,
+                     mime_type: nil, md5: nil, sha1: nil)
         @external_identifier = external_identifier
         @label = label
         @filename = filename
         @access = access
         @preserve = preserve
         @shelve = shelve
+        @mime_type = mime_type
+        @md5 = md5
+        @sha1 = sha1
       end
+      # rubocop:enable Metrics/ParameterLists
 
+      # rubocop:disable Metrics/MethodLength
       def as_json
         {
           "type": 'http://cocina.sul.stanford.edu/models/file.jsonld',
@@ -26,6 +34,26 @@ module SdrClient
             sdrPreserve: @preserve,
             shelve: @shelve
           }
+        }.tap do |json|
+          json['hasMessageDigests'] = message_digests unless message_digests.empty?
+          json['hasMimeType'] = @mime_type if @mime_type
+        end
+      end
+      # rubocop:enable Metrics/MethodLength
+
+      private
+
+      def message_digests
+        @message_digests ||= [].tap do |message_digests|
+          message_digests << create_message_digest('md5', @md5) unless @md5.nil?
+          message_digests << create_message_digest('sha1', @sha1) unless @sha1.nil?
+        end
+      end
+
+      def create_message_digest(algorithm, digest)
+        {
+          "type": algorithm,
+          digest: digest
         }
       end
     end

data/lib/sdr_client/deposit/file_set.rb

@@ -4,10 +4,11 @@ module SdrClient
   module Deposit
     # This represents the FileSet metadata that we send to the server for doing a deposit
     class FileSet
-      def initialize(uploads: [], files: [])
+      def initialize(uploads: [], uploads_metadata: {}, files: [], label:)
+        @label = label
         @files = if !uploads.empty?
                    uploads.map do |upload|
-                     File.new(external_identifier: upload.signed_id, label: upload.filename, filename: upload.filename)
+                     File.new(file_args(upload, uploads_metadata.fetch(upload.filename, {})))
                    end
                  else
                    files
@@ -17,15 +18,25 @@ module SdrClient
       def as_json
         {
           "type": 'http://cocina.sul.stanford.edu/models/fileset.jsonld',
+          "label": label,
           structural: {
-            hasMember: files.map(&:as_json)
+            contains: files.map(&:as_json)
           }
         }
       end
 
       private
 
-      attr_reader :files
+      attr_reader :files, :label
+
+      def file_args(upload, upload_metadata)
+        args = {
+          external_identifier: upload.signed_id,
+          label: upload.filename,
+          filename: upload.filename
+        }
+        args.merge(upload_metadata)
+      end
     end
   end
 end

data/lib/sdr_client/deposit/matching_file_grouping_strategy.rb

@@ -0,0 +1,17 @@
+# frozen_string_literal: true
+
+module SdrClient
+  module Deposit
+    # This strategy is for building one file set per set of similarly prefixed uploaded files
+    class MatchingFileGroupingStrategy
+      # @param [Array<SdrClient::Deposit::Files::DirectUploadResponse>] uploads the uploaded files to attach.
+      # @return [Array<Array<SdrClient::Deposit::Files::DirectUploadResponse>>] uploads the grouped uploaded files.
+      def self.run(uploads: [])
+        # Call `#values` on the result of the grouping operation because 1)
+        # `Process#build_filesets` expects an array of arrays, not an array of
+        # hashes, and 2) the keys aren't used anywhere
+        uploads.group_by { |ul| ::File.basename(ul.filename, '.*') }.values
+      end
+    end
+  end
+end

data/lib/sdr_client/deposit/process.rb

@@ -9,30 +9,39 @@ module SdrClient
       BLOB_PATH = '/v1/direct_uploads'
       DRO_PATH = '/v1/resources'
       # @param [Request] metadata information about the object
+      # @param [Class] grouping_strategy class whose run method groups an array of uploads
       # @param [String] url the server to send to
       # @param [String] token the bearer auth token for the server
       # @param [Array<String>] files a list of file names to upload
+      # @param [Hash<String, Hash<String, String>>] files_metadata file name, hash of additional file metadata
+      # Additional metadata includes access, preserve, shelve, md5, sha1
       # @param [Logger] logger the logger to use
-      def initialize(metadata:, url:, token:, files: [], logger: Logger.new(STDOUT))
+      # rubocop:disable Metrics/ParameterLists
+      def initialize(metadata:, grouping_strategy: SingleFileGroupingStrategy, url:,
+                     token:, files: [], files_metadata: {}, logger: Logger.new(STDOUT))
         @files = files
         @url = url
         @token = token
         @metadata = metadata
         @logger = logger
+        @grouping_strategy = grouping_strategy
+        @files_metadata = files_metadata
       end
+      # rubocop:enable Metrics/ParameterLists
 
       def run
         check_files_exist
         file_metadata = collect_file_metadata
         upload_responses = upload_file_metadata(file_metadata)
         upload_files(upload_responses)
-        request = metadata.with_uploads(upload_responses.values)
+        file_sets = build_filesets(uploads: upload_responses.values, files_metadata: files_metadata)
+        request = metadata.with_file_sets(file_sets)
         upload_metadata(request.as_json)
       end
 
       private
 
-      attr_reader :metadata, :files, :url, :token, :logger
+      attr_reader :metadata, :files, :url, :token, :logger, :grouping_strategy, :files_metadata
 
       def check_files_exist
         logger.info('checking to see if files exist')
@@ -111,6 +120,18 @@ module SdrClient
           conn.adapter :net_http
         end
       end
+
+      # @param [Array<SdrClient::Deposit::Files::DirectUploadResponse>] uploads the uploaded files to attach.
+      # @param [Hash<String,Hash<String, String>>] files_metadata filename, hash of additional file metadata.
+      # @return [Array<SdrClient::Deposit::FileSet>] the uploads transformed to filesets
+      def build_filesets(uploads:, files_metadata:)
+        grouped_uploads = grouping_strategy.run(uploads: uploads)
+        grouped_uploads.map.with_index(1) do |upload_group, i|
+          metadata_group = {}
+          upload_group.each { |upload| metadata_group[upload.filename] = files_metadata.fetch(upload.filename, {}) }
+          FileSet.new(uploads: upload_group, uploads_metadata: metadata_group, label: "Object #{i}")
+        end
+      end
     end
   end
 end

data/lib/sdr_client/deposit/request.rb

@@ -7,6 +7,7 @@ module SdrClient
       # @param [String] label the required object label
       # @param [String] type (http://cocina.sul.stanford.edu/models/object.jsonld) the required object type.
       # @param [Array<FileSet>] file_sets the file sets to attach.
+      # rubocop:disable Metrics/ParameterLists
       def initialize(label: nil,
                      apo:,
                      collection:,
@@ -22,6 +23,7 @@ module SdrClient
         @apo = apo
         @file_sets = file_sets
       end
+      # rubocop:enable Metrics/ParameterLists
 
       def as_json
         {
@@ -35,11 +37,8 @@ module SdrClient
         end
       end
 
-      # @param [Array<SdrClient::Deposit::Files::DirectUploadResponse>] uploads the uploaded files to attach.
-      # @return [Request] a clone of this request with the uploads added
-      def with_uploads(uploads)
-        file_sets = uploads.map { |upload| FileSet.new(uploads: [upload]) }
-
+      # @return [Request] a clone of this request with the file_sets added
+      def with_file_sets(file_sets)
         Request.new(label: label,
                     apo: apo,
                     collection: collection,
@@ -49,11 +48,6 @@ module SdrClient
                     file_sets: file_sets)
       end
 
-      # In this case there is a 1-1 mapping between Files and FileSets,
-      # but this doesn't always have to be the case.  We could change this in the
-      # future so that we have one FileSet that has an Image and its OCR file.
-      def add_uploads_each_as_resource(uploads); end
-
       private
 
       attr_reader :label, :file_sets, :source_id, :catkey, :apo, :collection, :type
@@ -73,7 +67,7 @@ module SdrClient
       def structural
         {
           isMemberOf: collection,
-          hasMember: file_sets.map(&:as_json)
+          contains: file_sets.map(&:as_json)
         }
       end
     end

data/lib/sdr_client/deposit/single_file_grouping_strategy.rb

@@ -0,0 +1,14 @@
+# frozen_string_literal: true
+
+module SdrClient
+  module Deposit
+    # This strategy is for building one file set per uploaded file
+    class SingleFileGroupingStrategy
+      # @param [Array<SdrClient::Deposit::Files::DirectUploadResponse>] uploads the uploaded files to attach.
+      # @return [Array<Array<SdrClient::Deposit::Files::DirectUploadResponse>>] uploads the grouped uploaded files.
+      def self.run(uploads: [])
+        uploads.map { |upload| [upload] }
+      end
+    end
+  end
+end

data/lib/sdr_client/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module SdrClient
-  VERSION = '0.3.0'
+  VERSION = '0.7.0'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: sdr-client
 version: !ruby/object:Gem::Version
-  version: 0.3.0
+  version: 0.7.0
 platform: ruby
 authors:
 - Justin Coyne
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2020-01-23 00:00:00.000000000 Z
+date: 2020-01-30 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: dry-monads
@@ -158,6 +158,7 @@ files:
 - bin/console
 - bin/setup
 - exe/sdr
+- lib/sdr-client.rb
 - lib/sdr_client.rb
 - lib/sdr_client/cli.rb
 - lib/sdr_client/credentials.rb
@@ -166,8 +167,10 @@ files:
 - lib/sdr_client/deposit/file_set.rb
 - lib/sdr_client/deposit/files/direct_upload_request.rb
 - lib/sdr_client/deposit/files/direct_upload_response.rb
+- lib/sdr_client/deposit/matching_file_grouping_strategy.rb
 - lib/sdr_client/deposit/process.rb
 - lib/sdr_client/deposit/request.rb
+- lib/sdr_client/deposit/single_file_grouping_strategy.rb
 - lib/sdr_client/login.rb
 - lib/sdr_client/login_prompt.rb
 - lib/sdr_client/version.rb
@@ -193,7 +196,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.3
+rubygems_version: 3.1.2
 signing_key: 
 specification_version: 4
 summary: The CLI for https://github.com/sul-dlss/sdr-api