moab-versioning 2.2.0 → 2.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 8a0b643706a4149747c76abbbbdb8458a73c62ed
-  data.tar.gz: 485960168209b4768a76deb35fbd56360a8b32e3
+  metadata.gz: e4b6638d612a715c99575b74595385e412fc8fd4
+  data.tar.gz: dec35bf65bb48c61e4f84447f06a4a244d556f8e
 SHA512:
-  metadata.gz: 26f4fee9182c007e5c68fdef45a53fbb75330dfcdc9ed39e02e719a095a1a4d684adbb2ab7ff049fa15631cf1e72711f78b3b6ee2633a99f31c264f2b77ddb05
-  data.tar.gz: 31c2bc1b8120546686d6952b06ab2b20b046d4f010a829bb546bc3c4669832ac15ef627b388dd4414e6c59e2628b1737f7a1c2dc1d07c37c3d26db9f365babea
+  metadata.gz: f1ae80fe9600cea7429abaeec22a74d962e7eec43738364f65860b5d127449bf22aae5cb25d237d17eeb469b918a155e200bce6d42f597c61ab1cc6a862e37c6
+  data.tar.gz: 2dd2b7cc61a003701e4bc2a08f7b63a0bd9b0927d03a30458ae24cc47e18b0e917709ffce7e6dcebbbd243ab5391456f84100b13d878e4ae1fcdd726348e6d47

data/lib/moab.rb

@@ -56,4 +56,3 @@ require 'moab/storage_repository'
 require 'moab/storage_services'
 require 'moab/exceptions'
 require 'moab/verification_result'
-

data/lib/moab/bagger.rb

@@ -192,7 +192,7 @@ module Moab
       manifest_file = Hash.new
       manifest_types =  [:md5, :sha1, :sha256]
       manifest_types.each do |type|
-        manifest_pathname[type] = @bag_pathname.join("manifest-#{type.to_s}.txt")
+        manifest_pathname[type] = @bag_pathname.join("manifest-#{type}.txt")
         manifest_file[type] = manifest_pathname[type].open('w')
       end
       @bag_inventory.groups.each do |group|
@@ -233,7 +233,7 @@ module Moab
       manifest_file = Hash.new
       manifest_types =  [:md5, :sha1, :sha256]
       manifest_types.each do |type|
-        manifest_pathname[type] = @bag_pathname.join("tagmanifest-#{type.to_s}.txt")
+        manifest_pathname[type] = @bag_pathname.join("tagmanifest-#{type}.txt")
         manifest_file[type] = manifest_pathname[type].open('w')
       end
       @bag_pathname.children.each do |file|

data/lib/moab/exceptions.rb

@@ -15,4 +15,4 @@ module Moab
 
   end
 
-end
+end

data/lib/moab/file_group.rb

@@ -219,12 +219,11 @@ module Moab
 
     # @api internal
     # @param pathname [Pathname, String] The location of the file to be added
-    # @param validated [Boolean] if true, path is verified to be descendant of (#base_directory)
+    # @param _validated (unused; kept here for backwards compatibility)
     # @return [void] Add a single physical file's data to the array of files in this group.
     #   If fixity data was supplied in bag manifests, then utilize that data.
-    def add_physical_file(pathname, validated=nil)
-      pathname=Pathname.new(pathname).expand_path
-      validated ||= is_descendent_of_base?(pathname)
+    def add_physical_file(pathname, _validated=nil)
+      pathname = Pathname.new(pathname).expand_path
       instance = FileInstance.new.instance_from_file(pathname, @base_directory)
       if @signatures_from_bag && @signatures_from_bag[pathname]
         signature = @signatures_from_bag[pathname]
@@ -234,7 +233,7 @@ module Moab
       else
         signature = FileSignature.new.signature_from_file(pathname)
       end
-      add_file_instance(signature,instance)
+      add_file_instance(signature, instance)
     end
 
   end

data/lib/moab/file_group_difference.rb

@@ -340,15 +340,15 @@ module Moab
       deltas = Hash.new {|hash, key| hash[key] = []}
       # case where other_path is empty or 'same'.  (create array of strings)
       [:identical, :modified, :deleted, :copydeleted].each do |change|
-        deltas[change].concat @subset_hash[change].files.collect{|file| file.basis_path}
+        deltas[change].concat(@subset_hash[change].files.collect{ |file| file.basis_path })
       end
       # case where basis_path and other_path are both present.  (create array of arrays)
       [:copyadded, :renamed].each do |change|
-        deltas[change].concat @subset_hash[change].files.collect{|file| [file.basis_path,file.other_path]}
+        deltas[change].concat(@subset_hash[change].files.collect { |file| [file.basis_path,file.other_path] })
       end
       # case where basis_path is empty.  (create array of strings)
       [:added].each do |change|
-        deltas[change].concat @subset_hash[change].files.collect{|file| file.other_path}
+        deltas[change].concat(@subset_hash[change].files.collect { |file| file.other_path })
       end
       deltas
     end

data/lib/moab/file_group_difference_subset.rb

@@ -42,4 +42,4 @@ module Moab
     has_many :files, FileInstanceDifference, :tag=>'file'
 
   end
-end
+end

data/lib/moab/file_inventory.rb

@@ -164,13 +164,13 @@ module Moab
       data_source = (groups.collect { |g| g.data_source.to_s }).join('|')
       if data_source.start_with?('contentMetadata')
         if version_id
-          "v#{version_id.to_s}-#{data_source}"
+          "v#{version_id}-#{data_source}"
         else
           "new-#{data_source}"
         end
       else
         if version_id
-          "v#{version_id.to_s}"
+          "v#{version_id}"
         else
           data_source
         end
@@ -214,7 +214,7 @@ module Moab
       manifest_pathname = Hash.new
       checksum_types =  [:md5, :sha1, :sha256]
       checksum_types.each do |type|
-        manifest_pathname[type] = bag_pathname.join("manifest-#{type.to_s}.txt")
+        manifest_pathname[type] = bag_pathname.join("manifest-#{type}.txt")
       end
       signatures = Hash.new { |hash,path| hash[path] = FileSignature.new }
       checksum_types.each do |type|
@@ -264,7 +264,7 @@ module Moab
         when "directory"
           'directoryInventory.xml'
         else
-          raise ArgumentError, "unknown inventory type: #{type.to_s}"
+          raise ArgumentError, "unknown inventory type: #{type}"
       end
     end
 

data/lib/moab/file_inventory_difference.rb

@@ -103,7 +103,7 @@ module Moab
     # @return [String] Returns either the common digitial object ID, or a concatenation of both inventory's IDs
     def common_object_id(basis_inventory, other_inventory)
       if basis_inventory.digital_object_id != other_inventory.digital_object_id
-        "#{basis_inventory.digital_object_id.to_s}|#{other_inventory.digital_object_id.to_s}"
+        "#{basis_inventory.digital_object_id}|#{other_inventory.digital_object_id}"
       else
         basis_inventory.digital_object_id.to_s
       end

data/lib/moab/file_signature.rb

@@ -79,7 +79,7 @@ module Moab
         when :sha256
           @sha256 = value
         else
-          raise ArgumentError, "Unknown checksum type '#{type.to_s}'"
+          raise ArgumentError, "Unknown checksum type '#{type}'"
       end
     end
 
@@ -89,7 +89,7 @@ module Moab
       checksum_hash[:md5] = @md5
       checksum_hash[:sha1] = @sha1
       checksum_hash[:sha256] = @sha256
-      checksum_hash.delete_if { |key,value| value.nil? or value.empty?}
+      checksum_hash.delete_if { |_key, value| value.nil? or value.empty? }
       checksum_hash
     end
 

data/lib/moab/signature_catalog.rb

@@ -136,7 +136,7 @@ module Moab
       group.files.each do |file|
         unless file.signature.complete?
           if @signature_hash.has_key?(file.signature)
-            file.signature = @signature_hash.find {|k,v| k == file.signature}[0]
+            file.signature = @signature_hash.find { |k, _v| k == file.signature }[0]
           elsif group_pathname
             file_pathname = group_pathname.join(file.instances[0].path)
             file.signature = file.signature.normalized_signature(file_pathname)

data/lib/moab/stanford.rb

@@ -4,6 +4,7 @@ require 'stanford/dor_metadata'
 require 'stanford/storage_repository'
 require 'stanford/storage_services'
 require 'stanford/active_fedora_object'
+require 'stanford/moab_storage_directory'
 
 # Stanford is a module that isolates classes specific to the Stanford Digital Repository
 #
@@ -14,4 +15,4 @@ require 'stanford/active_fedora_object'
 # @note Copyright (c) 2012 by The Board of Trustees of the Leland Stanford Junior University.
 #   All rights reserved.  See {file:LICENSE.rdoc} for details.
 module Stanford
-end
+end

data/lib/moab/storage_object.rb

@@ -200,7 +200,7 @@ module Moab
       end
     end
 
-    # @return [VerificationResult] Return result of storage verfication
+    # @return [VerificationResult] Return result of storage verification
     def verify_object_storage
       result = VerificationResult.new(digital_object_id)
       self.version_list.each do |version|

data/lib/moab/storage_object_version.rb

@@ -35,7 +35,7 @@ module Moab
     def initialize(storage_object, version_id)
       if version_id.is_a?(Integer)
         @version_id = version_id
-      elsif version_id.is_a?(String) and version_id.match /^v(\d+)$/
+      elsif version_id.is_a?(String) and version_id =~ /^v(\d+)$/
         @version_id = version_id.sub(/^v/,'').to_i
       else
         raise "version_id (#{version_id}) is not in a recognized format"
@@ -82,10 +82,10 @@ module Moab
       @storage_object.storage_filepath(catalog_filepath)
     end
 
-  # @param [String] file_category The category of file ('content', 'metadata', or 'manifest')
+  # @param _file_category (unused; kept here for backwards compatibility)
   # @param [FileSignature] file_signature The signature of the file
   # @return [Pathname] Pathname object containing the full path for the specified file
-    def find_filepath_using_signature(file_category, file_signature)
+    def find_filepath_using_signature(_file_category, file_signature)
       catalog_filepath = signature_catalog.catalog_filepath(file_signature)
       @storage_object.storage_filepath(catalog_filepath)
     end
@@ -140,7 +140,7 @@ module Moab
     # @param bag_dir [Pathname,String] The location of the bag to be ingested
     # @return [void] Create the version subdirectory and move files into it
     def ingest_bag_data(bag_dir)
-      raise "Version already exists: #{@version_pathname.to_s}" if @version_pathname.exist?
+      raise "Version already exists: #{@version_pathname}" if @version_pathname.exist?
       @version_pathname.join('manifests').mkpath
       bag_dir=Pathname(bag_dir)
       ingest_dir(bag_dir.join('data'),@version_pathname.join('data'))
@@ -225,9 +225,9 @@ module Moab
       result = VerificationResult.new("manifest_inventory")
       manifest_inventory = self.file_inventory('manifests')
       result.subentities << VerificationResult.verify_value('composite_key',self.composite_key,manifest_inventory.composite_key)
-      result.subentities << VerificationResult.verify_truth('manifests_group', ! manifest_inventory.group_empty?('manifests'))
+      result.subentities << VerificationResult.verify_truth('manifests_group', !manifest_inventory.group_empty?('manifests'))
       # measure the manifest signatures of the files in the directory (excluding manifestInventory.xml)
-      directory_inventory = FileInventory.new.inventory_from_directory(@version_pathname.join('manifests'),'manifests')
+      directory_inventory = FileInventory.new.inventory_from_directory(@version_pathname.join('manifests'), 'manifests')
       directory_inventory.digital_object_id = storage_object.digital_object_id
       directory_group = directory_inventory.group('manifests')
       directory_group.remove_file_having_path("manifestInventory.xml")
@@ -283,7 +283,7 @@ module Moab
           file.instances.each do |instance|
             relative_path = File.join(group.group_id, instance.path)
             catalog_entry = signature_catalog.signature_hash[file.signature]
-            if ! catalog_entry.nil?
+            if !catalog_entry.nil?
               found += 1
             else
               missing << relative_path.to_s

data/lib/moab/storage_repository.rb

@@ -123,8 +123,8 @@ module Moab
     unless storage_object.object_pathname.exist?
       if create
           storage_object.object_pathname.mkpath
-        else
-          raise Moab::ObjectNotFoundException, "No storage object found for #{object_id}"
+      else
+        raise Moab::ObjectNotFoundException, "No storage object found for #{object_id}"
       end
     end
     storage_object

data/lib/moab/storage_services.rb

@@ -139,7 +139,5 @@ module Moab
       compare_inventory=compare_version.file_inventory('version')
       FileInventoryDifference.new.compare(base_inventory,compare_inventory)
     end
-
   end
-
-end
+end

data/lib/moab/version_metadata_entry.rb

@@ -1,4 +1,4 @@
-  require 'moab'
+require 'moab'
 
 module Moab
 
@@ -61,4 +61,3 @@ module Moab
   end
 
 end
-

data/lib/serializer.rb

@@ -26,4 +26,3 @@ require 'digest/sha1'
 
 require 'serializer/serializable'
 require 'serializer/manifest'
-

data/lib/stanford/active_fedora_object.rb

@@ -6,7 +6,7 @@ module Stanford
   #
   # ====Data Model
   # * {DorMetadata} = utility methods for interfacing with Stanford metadata files (esp contentMetadata)
-  #   * {ContentInventory} [1..1] = utilities for transforming contentMetadata to versionInventory and doing comparsions
+  #   * {ContentInventory} [1..1] = utilities for transforming contentMetadata to versionInventory and doing comparisons
   #   * <b>{ActiveFedoraObject} [1..*] = utility for extracting content or other information from a Fedora Instance</b>
   #
   # @note Copyright (c) 2012 by The Board of Trustees of the Leland Stanford Junior University.

data/lib/stanford/content_inventory.rb

@@ -2,11 +2,11 @@ require 'moab/stanford'
 
 module Stanford
 
-  # Stanford-specific utility methods for transforming contentMetadata to versionInventory and doing
+  # Stanford-specific utility methods for transforming contentMetadata to versionInventory and doing comparisons
   #
   # ====Data Model
   # * {DorMetadata} = utility methods for interfacing with Stanford metadata files (esp contentMetadata)
-  #   * <b>{ContentInventory} [1..1] = utilities for transforming contentMetadata to versionInventory and doing comparsions</b>
+  #   * <b>{ContentInventory} [1..1] = utilities for transforming contentMetadata to versionInventory and doing comparisons</b>
   #   * {ActiveFedoraObject} [1..*] = utility for extracting content or other information from a Fedora Instance
   #
   # @note Copyright (c) 2012 by The Board of Trustees of the Leland Stanford Junior University.
@@ -97,7 +97,7 @@ module Stanford
     def generate_content_metadata(file_group, object_id, version_id)
       cm = Nokogiri::XML::Builder.new do |xml|
         xml.contentMetadata(:type=>"sample", :objectId=>object_id) {
-          xml.resource(:type=>"version", :sequence=>"1", :id=>"version-#{version_id.to_s}") {
+          xml.resource(:type=>"version", :sequence=>"1", :id=>"version-#{version_id}") {
             file_group.files.each do |file_manifestation|
               signature = file_manifestation.signature
               file_manifestation.instances.each do |instance|
@@ -122,7 +122,7 @@ module Stanford
     end
 
     # @param content_metadata [String,Nokogiri::XML::Document] The contentMetadata as a string or XML doc
-    # @return [Boolean] True if contentMetadata has essetial file attributes, else raise exception
+    # @return [Boolean] True if contentMetadata has essential file attributes, else raise exception
     def validate_content_metadata(content_metadata)
       result = validate_content_metadata_details(content_metadata)
       raise Moab::InvalidMetadataException, result[0]+" ..." if result.size > 0
@@ -197,7 +197,7 @@ module Stanford
       if file_size.nil? or file_size.empty?
         file_node['size'] = signature.size.to_s
       elsif file_size != signature.size.to_s
-        raise "Inconsistent size for #{file_node['id']}: #{file_size} != #{signature.size.to_s}"
+        raise "Inconsistent size for #{file_node['id']}: #{file_size} != #{signature.size}"
       end
     end
 
@@ -227,7 +227,7 @@ module Stanford
         if cm_checksum.nil? or cm_checksum.empty?
           checksum_node.content = sig_checksum
         elsif cm_checksum != sig_checksum
-          raise "Inconsistent #{type.to_s} for #{file_node['id']}: #{cm_checksum} != #{sig_checksum}"
+          raise "Inconsistent #{type} for #{file_node['id']}: #{cm_checksum} != #{sig_checksum}"
         end
       end
     end

data/lib/stanford/dor_metadata.rb

@@ -6,7 +6,7 @@ module Stanford
   #
   # ====Data Model
   # * <b>{DorMetadata} = utility methods for interfacing with Stanford metadata files (esp contentMetadata)</b>
-  #   * {ContentInventory} [1..1] = utilities for transforming contentMetadata to versionInventory and doing comparsions
+  #   * {ContentInventory} [1..1] = utilities for transforming contentMetadata to versionInventory and doing comparisons
   #   * {ActiveFedoraObject} [1..*] = utility for extracting content or other information from a Fedora Instance
   #
   # @note Copyright (c) 2012 by The Board of Trustees of the Leland Stanford Junior University.

data/lib/stanford/moab_storage_directory.rb

@@ -0,0 +1,36 @@
+require 'find'
+
+module Stanford
+  ##
+  # methods for dealing with a directory which stores Moab objects
+  class MoabStorageDirectory
+    DRUID_TREE_REGEXP = '[[:lower:]]{2}/\\d{3}/[[:lower:]]{2}/\\d{4}'.freeze
+    DRUID_REGEXP = '[[:lower:]]{2}\\d{3}[[:lower:]]{2}\\d{4}'.freeze
+
+    def self.find_moab_paths(storage_dir)
+      Find.find(storage_dir) do |path|
+        Find.prune unless File.directory?(path) # don't bother with a matching on files, we only care about directories
+        path_match_data = storage_dir_regexp(storage_dir).match(path)
+        if path_match_data
+          yield path_match_data[1], path, path_match_data # yield the druid, the full path, and the MatchData object
+          Find.prune # we don't care about what's in the moab dir, we just want the paths that look like moabs
+        end
+      end
+    end
+
+    def self.list_moab_druids(storage_dir)
+      druids = []
+      find_moab_paths(storage_dir) { |druid, _path, _path_match_data| druids << druid }
+      druids
+    end
+
+    private_class_method def self.storage_dir_regexps
+      @storage_dir_regexps ||= {}
+    end
+
+    # this regexp caching makes things many times faster (e.g. went from ~2200 s to crawl disk11, down to ~300 s)
+    private_class_method def self.storage_dir_regexp(storage_dir)
+      storage_dir_regexps[storage_dir] ||= Regexp.new("^#{storage_dir}/#{DRUID_TREE_REGEXP}/(#{DRUID_REGEXP})$")
+    end
+  end
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: moab-versioning
 version: !ruby/object:Gem::Version
-  version: 2.2.0
+  version: 2.3.0
 platform: ruby
 authors:
 - Darren Weber
@@ -11,7 +11,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-10-20 00:00:00.000000000 Z
+date: 2017-11-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: confstruct
@@ -196,7 +196,7 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
-  name: pry
+  name: pry-byebug
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
@@ -275,6 +275,7 @@ files:
 - lib/stanford/active_fedora_object.rb
 - lib/stanford/content_inventory.rb
 - lib/stanford/dor_metadata.rb
+- lib/stanford/moab_storage_directory.rb
 - lib/stanford/storage_repository.rb
 - lib/stanford/storage_services.rb
 - lib/tasks/yard.rake