Merge 50585be into 77b9265

lib/moab/storage_repository.rb

@@ -93,10 +93,33 @@ def find_storage_root(object_id, include_deposit = false)
     # @return [StorageObject] The representation of a digitial object's  storage directory, which might not exist yet.
     def find_storage_object(object_id, include_deposit = false)
       root = find_storage_root(object_id, include_deposit)
-      storage_pathname = root.join(storage_trunk, storage_branch(object_id))
-      storage_object = StorageObject.new(object_id, storage_pathname)
-      storage_object.storage_root = root
-      storage_object
+      create_storage_object(object_id, root)
+    end
+
+    # @param object_id [String] The identifier of the digital object
+    # @param include_deposit [Boolean] specifies whether to look in deposit areas for objects in process of initial ingest
+    # @return [Array<StorageObject>] Representations of a digitial object's storage directories, or an empty array if none found.
+    def search_storage_objects(object_id, include_deposit = false)
+      storage_objects = []
+      # Search for the object's home directory in the storage areas
+      branch = storage_branch(object_id)
+      storage_roots.each do |root|
+        root_trunk = root.join(storage_trunk)
+        raise(MoabRuntimeError, "Storage area not found at #{root_trunk}") unless root_trunk.exist?
+        root_trunk_branch = root_trunk.join(branch)
+        storage_objects << create_storage_object(object_id, root) if root_trunk_branch.exist?
+      end
+      # Search for the object's directory in the deposit areas
+      if include_deposit && deposit_trunk
+        branch = deposit_branch(object_id)
+        storage_roots.each do |root|
+          root_trunk = root.join(deposit_trunk)
+          raise(MoabRuntimeError, "Deposit area not found at #{root_trunk}") unless root_trunk.exist?
+          root_trunk_branch = root_trunk.join(branch)
+          storage_objects << create_storage_object(object_id, root) if root_trunk_branch.exist?
+        end
+      end
+      storage_objects
     end
 
     # @param object_id [String] The identifier of the digital object whose size is desired
@@ -129,5 +152,14 @@ def storage_object(object_id, create = false)
     def store_new_object_version(druid, bag_pathname)
       storage_object(druid, create = true).ingest_bag(bag_pathname)
     end
+
+    private
+
+    def create_storage_object(object_id, root)
+      storage_pathname = root.join(storage_trunk, storage_branch(object_id))
+      storage_object = StorageObject.new(object_id, storage_pathname)
+      storage_object.storage_root = root
+      storage_object
+    end
   end
 end

lib/moab/storage_services.rb

@@ -41,6 +41,13 @@ def self.find_storage_object(object_id, include_deposit = false)
       @@repository.find_storage_object(object_id, include_deposit)
     end
 
+    # @param object_id [String] The identifier of the digital object
+    # @param [Object] include_deposit
+    # @return [Array<StorageObject>] Representations of a digitial object's storage directories, or an empty array if none found.
+    def self.search_storage_objects(object_id, include_deposit = false)
+      @@repository.search_storage_objects(object_id, include_deposit)
+    end
+
     # @param object_id [String] The identifier of the digital object whose size is desired
     # @param include_deposit [Boolean] specifies whether to look in deposit areas for objects in process of initial ingest
     # @return [Integer] the size occupied on disk by the storage object, in bytes.  this is the entire moab (all versions).

spec/fixtures/derivatives2/ingests/jq937jp0017/v0001/data/metadata/contentMetadata.xml

@@ -0,0 +1,29 @@
+<?xml version="1.0"?>
+<contentMetadata type="sample" objectId="druid:jq937jp0017">
+  <resource type="version" sequence="1" id="version-1">
+    <file datetime="2012-03-26T08:15:11-06:00" size="40873" id="title.jpg" shelve="yes" publish="yes" preserve="yes">
+      <checksum type="MD5">1a726cd7963bd6d3ceb10a8c353ec166</checksum>
+      <checksum type="SHA-1">583220e0572640abcd3ddd97393d224e8053a6ad</checksum>
+    </file>
+    <file datetime="2012-03-26T08:20:35-06:00" size="41981" id="intro-1.jpg" shelve="yes" publish="yes" preserve="yes">
+      <checksum type="MD5">915c0305bf50c55143f1506295dc122c</checksum>
+      <checksum type="SHA-1">60448956fbe069979fce6a6e55dba4ce1f915178</checksum>
+    </file>
+    <file datetime="2012-03-26T08:19:30-06:00" size="39850" id="intro-2.jpg" shelve="yes" publish="yes" preserve="yes">
+      <checksum type="MD5">77f1a4efdcea6a476505df9b9fba82a7</checksum>
+      <checksum type="SHA-1">a49ae3f3771d99ceea13ec825c9c2b73fc1a9915</checksum>
+    </file>
+    <file datetime="2012-03-26T09:59:14-06:00" size="25153" id="page-1.jpg" shelve="yes" publish="yes" preserve="yes">
+      <checksum type="MD5">3dee12fb4f1c28351c7482b76ff76ae4</checksum>
+      <checksum type="SHA-1">906c1314f3ab344563acbbbe2c7930f08429e35b</checksum>
+    </file>
+    <file datetime="2012-03-26T09:23:36-06:00" size="39450" id="page-2.jpg" shelve="yes" publish="yes" preserve="yes">
+      <checksum type="MD5">82fc107c88446a3119a51a8663d1e955</checksum>
+      <checksum type="SHA-1">d0857baa307a2e9efff42467b5abd4e1cf40fcd5</checksum>
+    </file>
+    <file datetime="2012-03-26T09:24:39-06:00" size="19125" id="page-3.jpg" shelve="yes" publish="yes" preserve="yes">
+      <checksum type="MD5">a5099878de7e2e064432d6df44ca8827</checksum>
+      <checksum type="SHA-1">c0ccac433cf02a6cee89c14f9ba6072a184447a2</checksum>
+    </file>
+  </resource>
+</contentMetadata>

spec/fixtures/derivatives2/ingests/jq937jp0017/v0001/data/metadata/descMetadata.xml

@@ -0,0 +1,60 @@
+<?xml version="1.0"?>
+<mods:mods xmlns:mods="http://www.loc.gov/mods/v0003" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" version="3.3" xsi:schemaLocation="http://www.loc.gov/mods/v0003 http://cosimo.stanford.edu/standards/mods/v0003/mods-3-3.xsd">
+   <mods:titleInfo>
+      <mods:nonSort>The</mods:nonSort>
+      <mods:title>poems of Robert Burns</mods:title>
+      <mods:subTitle>A new edition</mods:subTitle>
+   </mods:titleInfo>
+   <mods:name type="personal">
+      <mods:namePart>Burns, Robert</mods:namePart>
+      <mods:namePart type="date">1759-1796</mods:namePart>
+      <mods:role>
+         <mods:roleTerm authority="marcrelator" type="text">creator</mods:roleTerm>
+      </mods:role>
+   </mods:name>
+   <mods:typeOfResource>text</mods:typeOfResource>
+   <mods:originInfo>
+      <mods:place>
+         <mods:placeTerm type="code" authority="marccountry">stk</mods:placeTerm>
+      </mods:place>
+      <mods:place>
+         <mods:placeTerm type="text">Edinburgh</mods:placeTerm>
+      </mods:place>
+      <mods:publisher>Printed by Oliver &amp; Co.</mods:publisher>
+      <mods:dateIssued>1806</mods:dateIssued>
+      <mods:dateIssued encoding="marc" keyDate="yes">1806</mods:dateIssued>
+      <mods:issuance>monographic</mods:issuance>
+   </mods:originInfo>
+   <mods:language>
+      <mods:languageTerm authority="iso639-2b" type="code">eng</mods:languageTerm>
+   </mods:language>
+   <mods:relatedItem type="original">
+      <mods:physicalDescription>
+         <mods:form authority="marcform">print</mods:form>
+         <mods:extent>1 p.  ., [5]-196 p. 12 cm.</mods:extent>
+      </mods:physicalDescription>
+      <mods:recordInfo>
+         <mods:recordContentSource authority="marcorg"/>
+         <mods:recordCreationDate encoding="marc">910108</mods:recordCreationDate>
+         <mods:recordChangeDate encoding="iso8601">19910108124826.0</mods:recordChangeDate>
+         <mods:recordIdentifier source="SUL catalog key">587254</mods:recordIdentifier>
+         <mods:recordIdentifier source="oclc">22934044</mods:recordIdentifier>
+      </mods:recordInfo>
+   </mods:relatedItem>
+   <mods:physicalDescription>
+      <mods:form authority="marcform">electronic</mods:form>
+      <mods:reformattingQuality>preservation</mods:reformattingQuality>
+      <mods:digitalOrigin>reformatted digital</mods:digitalOrigin>
+   </mods:physicalDescription>
+   <mods:note>Title vignette.</mods:note>
+   <mods:recordInfo>
+      <mods:recordContentSource>DOR_MARC2MODS3-3.xsl Revision 1.1</mods:recordContentSource>
+      <mods:recordCreationDate encoding="iso8601">2011-08-25T08:39:54.573-07:00</mods:recordCreationDate>
+      <mods:recordIdentifier source="Data Provider Digital Object Identifier">36105024276136</mods:recordIdentifier>
+   </mods:recordInfo>
+   <mods:identifier type="local" displayLabel="SUL Resource ID">druid:jq937jp0017</mods:identifier>
+   <mods:location>
+      <mods:physicalLocation>Stanford University Libraries</mods:physicalLocation>
+      <mods:url>http://purl.stanford.edu/jq937jp0017</mods:url>
+   </mods:location>
+</mods:mods>

spec/fixtures/derivatives2/ingests/jq937jp0017/v0001/data/metadata/identityMetadata.xml

@@ -0,0 +1,20 @@
+<identityMetadata>
+  <objectAdminClass>GoogleBooks</objectAdminClass>
+  <agreementId>druid:zn292gq7284</agreementId>
+  <objectType>item</objectType>
+  <citationTitle>The poems of Robert Burns: A new edition</citationTitle>
+  <citationCreator>Burns, Robert , 1759-1796</citationCreator>
+  <otherId name="catkey">587254</otherId>
+  <otherId name="shelfseq">821.5 .B961O T</otherId>
+  <otherId name="barcode">36105024276136</otherId>
+  <otherId name="callseq">1</otherId>
+  <otherId name="uuid">89abf2c4-c121-41f0-8838-484a661c92d9</otherId>
+  <objectId>druid:jq937jp0017</objectId>
+  <objectCreator>DOR</objectCreator>
+  <sourceId source="google">STANFORD_36105024276136</sourceId>
+  <objectLabel>Google Scanned Book, barcode 36105024276136</objectLabel>
+  <tag>Google Book : GBS VIEW_FULL</tag>
+  <tag>Book : Non-US pre-1891</tag>
+  <tag>Google Book : Scan source STANFORD</tag>
+  <tag>Remediated By : 3.6.2</tag>
+</identityMetadata>

spec/fixtures/derivatives2/ingests/jq937jp0017/v0001/data/metadata/provenanceMetadata.xml

@@ -0,0 +1,86 @@
+<provenanceMetadata xmlns:PREMIS="info:lc/xmlns/premis-v2" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" objectId="druid:jq937jp0017">
+  <agent name="Google">
+    <PREMIS:premis version="2.0">
+          <PREMIS:object xsi:type="representation">
+            <PREMIS:objectIdentifier>
+              <PREMIS:objectIdentifierType>barcode</PREMIS:objectIdentifierType>
+              <PREMIS:objectIdentifierValue>STANFORD_36105024276136</PREMIS:objectIdentifierValue>
+            </PREMIS:objectIdentifier>
+            <PREMIS:significantProperties>
+              <PREMIS:significantPropertiesType>file count</PREMIS:significantPropertiesType>
+              <PREMIS:significantPropertiesValue>410</PREMIS:significantPropertiesValue>
+            </PREMIS:significantProperties>
+            <PREMIS:significantProperties>
+              <PREMIS:significantPropertiesType>page count</PREMIS:significantPropertiesType>
+              <PREMIS:significantPropertiesValue>204</PREMIS:significantPropertiesValue>
+            </PREMIS:significantProperties>
+          </PREMIS:object>
+          <PREMIS:event xmlID="Google_Scan">
+            <PREMIS:eventIdentifier>
+              <PREMIS:eventIdentifierType>Google generated</PREMIS:eventIdentifierType>
+              <PREMIS:eventIdentifierValue>Google_Scan</PREMIS:eventIdentifierValue>
+            </PREMIS:eventIdentifier>
+            <PREMIS:eventType>Scan</PREMIS:eventType>
+            <PREMIS:eventDateTime>2009-03-25T07:04:20</PREMIS:eventDateTime>
+            <PREMIS:eventDetail>Original image capture</PREMIS:eventDetail>
+            <PREMIS:linkingAgentIdentifier>
+              <PREMIS:linkingAgentIdentifierType>AgentID</PREMIS:linkingAgentIdentifierType>
+              <PREMIS:linkingAgentIdentifierValue>Google (Firm)</PREMIS:linkingAgentIdentifierValue>
+            </PREMIS:linkingAgentIdentifier>
+          </PREMIS:event>
+          <PREMIS:event xmlID="Google_Process">
+            <PREMIS:eventIdentifier>
+              <PREMIS:eventIdentifierType>Google generated</PREMIS:eventIdentifierType>
+              <PREMIS:eventIdentifierValue>Google_Process</PREMIS:eventIdentifierValue>
+            </PREMIS:eventIdentifier>
+            <PREMIS:eventType>Process</PREMIS:eventType>
+            <PREMIS:eventDateTime>2010-04-04T06:24:51</PREMIS:eventDateTime>
+            <PREMIS:eventDetail>Refinement and cleaning of raw scanned images using image enhancement techniques</PREMIS:eventDetail>
+            <PREMIS:linkingAgentIdentifier>
+              <PREMIS:linkingAgentIdentifierType>AgentID</PREMIS:linkingAgentIdentifierType>
+              <PREMIS:linkingAgentIdentifierValue>Google (Firm)</PREMIS:linkingAgentIdentifierValue>
+            </PREMIS:linkingAgentIdentifier>
+          </PREMIS:event>
+          <PREMIS:event xmlID="Google_Analyze">
+            <PREMIS:eventIdentifier>
+              <PREMIS:eventIdentifierType>Google generated</PREMIS:eventIdentifierType>
+              <PREMIS:eventIdentifierValue>Google_Analyze</PREMIS:eventIdentifierValue>
+            </PREMIS:eventIdentifier>
+            <PREMIS:eventType>Analyze</PREMIS:eventType>
+            <PREMIS:eventDateTime>2011-02-01T13:59:22</PREMIS:eventDateTime>
+            <PREMIS:eventDetail>Automated quality assessment, extraction of OCR text and metadata, assignment of metadata tags, and page ordering</PREMIS:eventDetail>
+            <PREMIS:linkingAgentIdentifier>
+              <PREMIS:linkingAgentIdentifierType>AgentID</PREMIS:linkingAgentIdentifierType>
+              <PREMIS:linkingAgentIdentifierValue>Google (Firm)</PREMIS:linkingAgentIdentifierValue>
+            </PREMIS:linkingAgentIdentifier>
+          </PREMIS:event>
+          <PREMIS:event xmlID="Google_Convert">
+            <PREMIS:eventIdentifier>
+              <PREMIS:eventIdentifierType>Google generated</PREMIS:eventIdentifierType>
+              <PREMIS:eventIdentifierValue>Google_Convert</PREMIS:eventIdentifierValue>
+            </PREMIS:eventIdentifier>
+            <PREMIS:eventType>Convert</PREMIS:eventType>
+            <PREMIS:eventDateTime>2011-08-25T08:38:33</PREMIS:eventDateTime>
+            <PREMIS:eventDetail>Creation of partner deliverables</PREMIS:eventDetail>
+            <PREMIS:linkingAgentIdentifier>
+              <PREMIS:linkingAgentIdentifierType>AgentID</PREMIS:linkingAgentIdentifierType>
+              <PREMIS:linkingAgentIdentifierValue>Google (Firm)</PREMIS:linkingAgentIdentifierValue>
+            </PREMIS:linkingAgentIdentifier>
+          </PREMIS:event>
+          <PREMIS:agent>
+            <PREMIS:agentIdentifier>
+              <PREMIS:agentIdentifierType>LCNAF</PREMIS:agentIdentifierType>
+              <PREMIS:agentIdentifierValue>Google (Firm)</PREMIS:agentIdentifierValue>
+            </PREMIS:agentIdentifier>
+            <PREMIS:agentType>organization</PREMIS:agentType>
+          </PREMIS:agent>
+        </PREMIS:premis>
+  </agent>
+  <agent name="DOR">
+  <what object="druid:jq937jp0017">
+    <event who="DOR-robot:process-content" when="2011-08-26T01:45:37-07:00">Image files JHOVE 1.4 validated</event>
+    <event who="DOR-robot:register-object" when="2011-08-25T08:21:22-0700">Google data received</event>
+    <event who="DOR-robot:google-download" when="2011-08-25T21:24:13-0700">Checksums verified</event>
+  </what>
+</agent>
+</provenanceMetadata>

spec/fixtures/derivatives2/ingests/jq937jp0017/v0001/data/metadata/versionMetadata.xml

@@ -0,0 +1,6 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<versionMetadata objectId="druid:jq937jp0017">
+  <version versionId="1" label="1.0" significance="major">
+    <description>Initial version</description>
+  </version>
+</versionMetadata>

spec/fixtures/derivatives2/ingests/jq937jp0017/v0001/manifests/fileInventoryDifference.xml

@@ -0,0 +1,56 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<fileInventoryDifference objectId="druid:jq937jp0017" differenceCount="11" basis="v0" other="v1" reportDatetime="2019-12-17T18:47:20Z">
+  <fileGroupDifference groupId="content" differenceCount="6" identical="0" copyadded="0" copydeleted="0" renamed="0" modified="0" added="6" deleted="0">
+    <subset change="added" count="6">
+      <file change="added" basisPath="" otherPath="intro-1.jpg">
+        <fileSignature size="41981" md5="915c0305bf50c55143f1506295dc122c" sha1="60448956fbe069979fce6a6e55dba4ce1f915178" sha256="4943c6ffdea7e33b74fd7918de900de60e9073148302b0ad1bf5df0e6cec032a"/>
+      </file>
+      <file change="added" basisPath="" otherPath="intro-2.jpg">
+        <fileSignature size="39850" md5="77f1a4efdcea6a476505df9b9fba82a7" sha1="a49ae3f3771d99ceea13ec825c9c2b73fc1a9915" sha256="3a28718a8867e4329cd0363a84aee1c614d0f11229a82e87c6c5072a6e1b15e7"/>
+      </file>
+      <file change="added" basisPath="" otherPath="page-1.jpg">
+        <fileSignature size="25153" md5="3dee12fb4f1c28351c7482b76ff76ae4" sha1="906c1314f3ab344563acbbbe2c7930f08429e35b" sha256="41aaf8598c9d8e3ee5d55efb9be11c542099d9f994b5935995d0abea231b8bad"/>
+      </file>
+      <file change="added" basisPath="" otherPath="page-2.jpg">
+        <fileSignature size="39450" md5="82fc107c88446a3119a51a8663d1e955" sha1="d0857baa307a2e9efff42467b5abd4e1cf40fcd5" sha256="235de16df4804858aefb7690baf593fb572d64bb6875ec522a4eea1f4189b5f0"/>
+      </file>
+      <file change="added" basisPath="" otherPath="page-3.jpg">
+        <fileSignature size="19125" md5="a5099878de7e2e064432d6df44ca8827" sha1="c0ccac433cf02a6cee89c14f9ba6072a184447a2" sha256="7bd120459eff0ecd21df94271e5c14771bfca5137d1dd74117b6a37123dfe271"/>
+      </file>
+      <file change="added" basisPath="" otherPath="title.jpg">
+        <fileSignature size="40873" md5="1a726cd7963bd6d3ceb10a8c353ec166" sha1="583220e0572640abcd3ddd97393d224e8053a6ad" sha256="8b0cee693a3cf93cf85220dd67c5dc017a7edcdb59cde8fa7b7f697be162b0c5"/>
+      </file>
+    </subset>
+    <subset change="identical" count="0"/>
+    <subset change="copyadded" count="0"/>
+    <subset change="copydeleted" count="0"/>
+    <subset change="renamed" count="0"/>
+    <subset change="modified" count="0"/>
+    <subset change="deleted" count="0"/>
+  </fileGroupDifference>
+  <fileGroupDifference groupId="metadata" differenceCount="5" identical="0" copyadded="0" copydeleted="0" renamed="0" modified="0" added="5" deleted="0">
+    <subset change="added" count="5">
+      <file change="added" basisPath="" otherPath="contentMetadata.xml">
+        <fileSignature size="1871" md5="8cdee9c3470552d258a4351bf4c117e2" sha1="321737e42cc4a3134164539f768514d4f7f6d184" sha256="af38d344d9aee248b01b0af9c85dffbfbd1aeb0f2c6dadf3620797ca73ab24c3"/>
+      </file>
+      <file change="added" basisPath="" otherPath="descMetadata.xml">
+        <fileSignature size="3055" md5="19c96e98dd25dd68c493faa6a0fde4d0" sha1="541d5845cc866d7676fa43c337b3f7a59f597487" sha256="a575d9058a56d77a4e65e2b1bedb619686a03d5f717374bd0df98d422432e1fe"/>
+      </file>
+      <file change="added" basisPath="" otherPath="identityMetadata.xml">
+        <fileSignature size="932" md5="f0815d7b45530491931d5897ccbe2dd1" sha1="4065ff5523e227c1914098372a3dc587f739030e" sha256="f48727374e73776d00517dcdc8af62ec8b18a24cff46c66a5c5f85515144628e"/>
+      </file>
+      <file change="added" basisPath="" otherPath="provenanceMetadata.xml">
+        <fileSignature size="5306" md5="17193dbf595571d728ba59aa31638db9" sha1="c8b91eacf9ad7532a42dc52b3c9cf03b4ad2c7f6" sha256="d38899a57d4cfca2b1ac73356c8a89cd3c902d35b853ae52f468004887b73dbe"/>
+      </file>
+      <file change="added" basisPath="" otherPath="versionMetadata.xml">
+        <fileSignature size="224" md5="36e3c1dadad827cb49e521f5d6559127" sha1="af96327b72bf37e3e60cf51c6de4dba1f6dea620" sha256="65426a7389e13dfa73d37c0ab1417c3e31de74289bf77cc74b3dcc3ffa6f4c8e"/>
+      </file>
+    </subset>
+    <subset change="identical" count="0"/>
+    <subset change="copyadded" count="0"/>
+    <subset change="copydeleted" count="0"/>
+    <subset change="renamed" count="0"/>
+    <subset change="modified" count="0"/>
+    <subset change="deleted" count="0"/>
+  </fileGroupDifference>
+</fileInventoryDifference>