Differences between revisions 1 and 90 (spanning 89 versions)
⇤ ← Revision 1 as of 2008-06-26 12:26:05 →
Size: 15678
Editor: kfc
Comment: Created by the PackagePages action.
← Revision 90 as of 2010-03-17 13:13:00 → ⇥
Size: 5322
Editor: localhost
Comment: converted to 1.6 markup
Deletions are marked like this. Additions are marked like this.
Line 1: Line 1:
= DOMS Data model = = DON'T PANIC =
Line 3: Line 3:
[[ImageLink(http://merkur/viewvc/trunk/docs/datamodel/fig/DOMSBaseCollection.png?root=doms&view=co,alt=DOMS base collection,width=1000)]] '''A definition: A ''datamodel'' describes the content of a collection. A ''content model'' describes the content of a data object. So, a datamodel is a set of content models, that together describe the collection.'''
Line 5: Line 5:
The DOMS datamodel describe how the Type system underlying DOMS is realised in Fedora 3. The figure above will serve as a guide through the following sections. The DOMS datamodel describes how the Type system underlying DOMS is realized in Fedora 3.

The entire DOMS datamodel is, in its entirety, a complex system. For proper understanding, the various components have been detailed in seperate documents. Firstly, it consist of a number of extensions to the Fedora system. Secondly, it consist of a number of predefined objects. These objects make use of the extensions to Fedora. Thirdly, it consist of a number of policies for how certain tasks are achived. And fourthly, it consist of a number of API interfaces.

Fedora and DOMS are big on namespaces. To ease writing the documentation, a namespace document, DomsNameSpacesAndSchemas, have been written. All namespaces should be defined there, and all shorthands refer to the namespaces defined therein.
Line 8: Line 12:
== Content Models in general ==
Line 9: Line 14:
== DOMS Content Models ==

Fedora provides a repository for digital objects. All objects in the repository can, in principle be unique, but Fedora provides of specifying an object as a given type. Unfortunately, the type-definitions in Fedora, called Content Models, are rather simplistic by default. We use them as the basis of our type system, with certain enhancements.		 Fedora provides a repository for digital objects. All objects in the repository can, in principle, be unique, but Fedora provides a way of specifying that an object has a given type. Unfortunately, the type-definitions in Fedora, called Content Models, are rather simplistic by default. We use them as the basis of our type system, with certain enhancements.
Line 17: Line 20:
The Content Model object, as used in DOMS, describes the compulsary and legal content of an object of this type. It contains the information nessesary to verify if the given object is indeed of this type. The Content Model object, as used in DOMS, describes the compulsory and legal content of an object of this type. It contains the information nessesary to verify if the given object is indeed of this type. For more detail on this, see FedoraOntology and FedoraTypeChecking
Line 19: Line 22:
A data object can specify the Content Model describing it's contents, via a hasContentModel relation. It can only have one such relation, and in DOMS we require it to be present. A data object can specify the Content Model describing its contents, via a fedora-model:hasModel relation, and in DOMS we require it to be present. A data object will be said to "subcribe" to a Content Model. Content Model inheritance, as specified in FedoraOntology, will be used.
Line 21: Line 24:
=== Inheritance ===
A type system that does not allow for inheritance, will have limited use. In spite of this, the Fedora Content Models do not provide this functionality. We have build our own inheritance system for Content Models to compensate for this lack.		 The special Content Model object "doms:!ContentModel_DOMS" is the root object. All Content Models must have an "doms-relations:extendsModel" relation to this object, possibly through a number of other Content Models. The complete description of a data object is defined as the set of the descriptions in the Content Model specified with "fedora-model:hasModel" and all Content Models that can be reached from this, by following "doms-relations:extendsModel" relations.
Line 24: Line 26:
The special Content Model object "ContentModel_DOMS" is the root object. All Content Models must have an "extends" relation to this object, possible through a number of other Content Models. When performing validation of a Data Object, the validator will validate the object against the Content Model specified with "hasContentModel" and all Content Models that can be reached from this, by following "extends" relations.

A Content Model can "extend" more than one other Content Model. The validator should only validate an object against a given Content Model once per invocation.

=== Predefined Content Models ===

Shorthand:
 * myObject.myDatastream means the Datastream myDatastream in the Object myObject.

==== ContentModel_DOMS ====

The following variables are used:
 * $Title: A human readable title
 * $PID: A generated PID for this object
 * $ContentModel: The content model for this object, must derive from ContentModel_DOMS
 * $Collection: The collection that this object belongs to. An object with a Content Model that derives from ContentModel_Collection
 * $Licence: The Licence object that govern access to this object. An object with a Content Model that derives from ContentModel_Licence

Requirements for objects described by ContentModel_DOMS
 * Datastreams
  * DC
   * dc:title = $Title
  * DomsDC
   * dcterms:title = $Title
  * RELS-EXT
   * oai:itemID = $PID
   * fedora:hasContentModel -> $ContentModel
   * doms:hasLicence -> $LicenceObject
   * doms:isPartOfCollection -> $Collection
  * AUDIT
   Systemcontrolled audit trail
  * POLICY
   * ContentLocation URL = $LicenceObject.LICENCE
  * STATE: The state of the object, in XML
   * <availibility> = draft | intermediate | published

==== ContentModel_Image ====
The following variables are used:
 * $TiffFile: An object with the Content Model ContentModel_TiffFile;

Requirements for objects described by ContentModel_Image
 * Datastreams
  * RELS-EXT
   * doms:hasFile -> $TiffFile

==== ContentModel_Audio ====
The following variables are used:
 * $WavFile: An object with the Content Model ContentModel_WavFile;
 * $BwfFile: An object with the Content Model ContentModel_BwfFile;

Requirements for objects described by ContentModel_Wav
 * Datastreams
  * RELS-EXT: One of the following
   * doms:hasFile -> $WavFile
   * doms:hasFile -> $BwfFile

==== ContentModel_Audio ====
The following variables are used:
 * $WavFile: An object with the Content Model ContentModel_WavFile;
 * $BwfFile: An object with the Content Model ContentModel_BwfFile;

Requirements for objects described by ContentModel_Wav
 * Datastreams
  * RELS-EXT: One of the following
   * doms:hasFile -> $WavFile
   * doms:hasFile -> $BwfFile

==== ContentModel_Video ====
The following variables are used:
 * $Mpeg1File: An object with the Content Model ContentModel_Mpeg1File;
 * $Mpeg2File: An object with the Content Model ContentModel_Mpeg2File;

Requirements for objects described by ContentModel_Wav
 * Datastreams
  * RELS-EXT: One of the following
   * doms:hasFile -> $Mpeg1File
   * doms:hasFile -> $Mpeg2File

==== ContentModel_Text ====
The following variables are used:
 * $Utf8File: An object with the Content Model ContentModel_WavFile;
 * $PdfFile: An object with the Content Model ContentModel_BwfFile;
 * $DocFile: An object with the Content Model ContentModel_OfficeOpenXmlFile;
Requirements for objects described by ContentModel_Wav
 * Datastreams
  * RELS-EXT: One of the following
   * doms:hasFile -> $Utf8File
   * doms:hasFile -> $PdfFile
   * doms:hasFile -> $DocFile		 A Content Model can "extend" more than one other Content Model. There is no overriding of Content Models, a subscribing object must be valid in regards to all the Content Models in the inheritance tree.
Line 115: Line 29:
==== ContentModel_Licence ====

Requirements for objects described by ContentModel_Licence
 * Datastreams
  * LICENCE: XACML describing the licence
  * LICENCETEXT: The human readable textual representation of the licence

==== ContentModel_Collection ====

Requirements for objects described by ContentModel_Collection: None		 Content Models have two datastreams in particular that are interesting. These are the ONTOLOGY and DS-COMPOSITE. The Ontology defines the the allowed relations in subscribing objects, and the DS-COMPOSITE defines the required datastreams and any restrictions they must adhere to.
Line 127: Line 32:
==== ContentModel_File ====
The following variables are used:
 * $OrigFile: An object with the Content Model that derives from ContentModel_File; 		== Fedora extensions ==
 * FedoraOntology - Detailing the Ontology lanquage
 * FedoraTypeChecking - Detailing the extensions to the DS-COMPOSITE stream
 * FedoraViewBlobs - Detailing the View system, which allows you to view many objects as one
 * FedoraObjectTemplates - Detailing the prototype system for content models
 * FedoraImportExport - Detailing use of Import and Export of objects using content models
 * FedoraLicensePolicies - Detailing the License system, and how it interacts with the Search system
 * FedoraState - Detailing the use of object states to allow controlling validity, availability and deletions.
Line 131: Line 41:
Requirements for objects described by ContentModel_File
 * Datastreams
  * RELS-EXT
   * (optional) doms:hasOriginal -> $OrigFile
  * CHARACTERIZATION: The output of the characterization tools
  * CONTENTS:
   * ContentLocation URL = The file in Bitstorage
  * ORIGIN: Metadata about the creation of the file
  * PRONOM: The pronom ID of the fileformat		 == Predefined objects ==
 * DomsPredefinedObjects - The predefined content models and other objects in the doms system
Line 141: Line 44:
All the predefined subtypes of File bring no new requirements. == Doms policies ==
 * DomsFileHandling - Detailing how we expect to handle file objects in Fedora.
 * DomsCollections - Detailing the use of collection objects in DOMS
 * DomsAuditTrail - Detailing how DOMS logs changes and new versions
Line 143: Line 49:
=== Content Model implementation ===

A Content Model in DOMS must have a number of additional datastreams, in regards to the Content Model requirements defined by Fedora.

The following variables are used:
 * $PID: A generated PID for this object
 * $ContentModel: The content model for this object, must derive from ContentModel_DOMS
 * $Collection: The collection that this object belongs to. An object with a Content Model that derives from ContentModel_Collection
 * $Licence: The Licence object that govern access to this object. An object with a Content Model that derives from ContentModel_Licence		 == API documentation ==
 * Refer to the overall [[Documentation]] page
Line 154: Line 53:
Requirements for Content Model objects (except ContentModel_DOMS, which do not have an "extends" relation)
 * Datastreams
  * RELS-EXT
   * (1+) doms:extends -> $ContentModel
   * oai:itemID = $PID
   * doms:hasLicence -> $LicenceObject
   * doms:isPartOfCollection -> $Collection
  * VALIDATIONBINDINGS: Described below
  * VIEW: Described below
  * AUDIT
   Systemcontrolled audit trail
  * POLICY
   * ContentLocation URL = $LicenceObject.LICENCE
  * STATE: The state of the object, in XML
   * <availibility> = draft | intermediate | published
  
Line 172: Line 55:
VALIDATIONBINDINGS contain xml of the form:
{{{
<binding name="binding1">
  <from name="datastream_name_in_data_object"/>
  <to name="datastream_with_validator_schema_in_this_object"/>
</binding>
}}}		 == Working with the Data model ==
Doms contains a number of content models. These are meant to serve as the basic buildingblocks for data models for new collections. A datamodel is, of course, not restricted to use only these content models, it can, and should, define it's own. All new content models, should extend doms:!ContentModel_DOMS, and all objects that need to reference files outside Fedora should have a content model that derive from doms:!ContentModel_File and so on. The content models that provide extra meaning are optional to use, and should at least be extended for the relevant collection.
Line 180: Line 58:
First stab at a schema for this datastream
{{{
<xsd:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
  <xsd:element name="bindings">
    <xsd:complexType>
      <xsd:sequence>
       <xsd:element name="binding" type="bindingType" minOccurs="0" maxOccurs="unbounded"/>
      </xsd:sequence>
    </xsd:complexType>
  <xsd:element>
  <xsd:complexType name="bindingType">
    <xsd:sequence>
      <xsd:element name="from" type="datastreamType"/>
      <xsd:element name="to" type="datastreamType"/>
   </xsd:sequence>
  </xsd:complexType>
  <xsd:complexType name="datastreamType">
    <xsd:attribute name="ID" type="idType" use="required"/>
  </xsd:complexType>
  <xsd:simpleType name="idType">
    <xsd:restriction base="xsd:ID">
      <xsd:maxLength value="64"/>
    </xsd:restriction>
  </xsd:simpleType>
</xs:schema>
}}}


== Collections ==

The DOMS system will be a system that models several collections of digital objects. Each object belongs to one or more collections. This is represented by having one or more "isPartOfCollection" relations to the parent collections. This goes for a collection object as well - it belongs to another collection. One collection has special status though: the "doms:Root_collection" does not belong to any other collection, and is thus the bottom element for the "isPartOfCollection" relation. Every other collection has a "isPartOfCollection" relation to "doms:Root_collection".

In addition, DOMS contains a special collection - the "doms:DOMS_base_collection". This collection provides objects such as content models and licences that are utilized by (and mandatory for) the other collections in the DOMS. Content models and licences are discussed below. This collection is meant to be ingested as the first collection in the DOMS.

== File Objects ==

In DOMS, we have found it beneficial to seperate the abstract concept of "Image" or "Audio" from the concrete implementations such as "jpeg" and "mp3". The metadata about the image will be relevant no matter the manifestation of the image, and as such should not reside along with the technical metadata about the manifestation. To support this separation, we have introduced the concept of File objects.

A File object is an object, that contain a link to the file (in Bitstorage), and the technical metadata about this file. Only File objects are allowed to reference a file in Bitstorage. File objects must all have a Content Model that extends ContentModel_File.

=== Preservation files ===
We have a special class of files in DOMS, the ones we are willing to promise to preserve. These are the eight formats
 * Tiff
 * Wav, Bwf
 * Mpeg1, Mpeg2
 * UTF8, Pdf, OfficeOpenXml
For each of these, we have defined a Content Model that extends ContentModel_File.

=== Presentation files ===
DOMS does not, by default, support the concept of presentation files. Files in Bitstorage, are per default, meant to preserved, and it should not be used for transitory formats. Instead presentation files are dynamically generated upon request.

All the preservation files Content Models have disseminators that convert the given preservation format into a presentation format.

=== Technical Metadata ===
A file object should contain technical metadata. In this context it refers
 * A datastream with the output of the characterization tools used on this file upon ingest
 * A datastream with the metadata about the origins of the file.

In addition, it must have a relation "hasOriginal" if it was migrated from another file that exist in DOMS.


== Views ==

A view is a way of combining objects in the DOMS into a domain-relevant group. It is a way of seeing a number of objects as related - as a whole; information that can be useful for the GUI-generator when generating GUI-windows.

Those views that we imagine as being suitable for a screen or window in the GUI, are called main views. Each main view contains an object that the main view is centered around. We call this the main object, and the ID of this view is the ID of the main object. Views of other objects are simply called views.
The main object is the object that represents the main view - other objects in that view are related to the main object and would presumably be relevant to edit in the GUI at the same time. For a CD modelled in DOMS, for example, a CD object would be the main object, and objects for tracks, cover, lyrics and so on would constitute the rest of the main view.

We imagine that results appearing in searches in the GUI will all be main views. In fact every view that will be the basis for a screen/window will be a main view.

A view for an object O is represented by a Datastream VIEW on the Content Model object for O. This Datastream also mark the object as Main, if this is the case. Please note that the view is defined on Content Model level, so the same rules are used to generate the view for all objects using that Content Model.

The datastream will just contain a list of relation names. Following these relations will give you the view. Views are inherited when Content Models "extends" each other, so you should generate the view for each Content Model in the inheritance tree of this Object, and remove duplicates.

Note: Exactly how these relations should be followed has not been decided yet. Suggestions include:

 * 1-step relations (relations on a content model c of the form "x rel y", meaning that if an object x with content model c has relation rel to another object y, then y will be part the view too. Examples of rel for a CD modelled in DOMS could be hasTrack, hasLyrics,..)

 * x-step relations. These are relations of the form mentioned above, but they will be followed from an object an arbitrary number of times, as long as the relations match.

 * reverse relations (relations of the type "y rel x" on a content model c, where x has model c and y therefore will be included in the view for x)

In addition, it has been suggested to augment the 1-step approach with the idea of "includes". What this means is that when object O has a view defined by following relations from O once, and an object P is in the view of O, then the view of P will be included in the view of O.
Note that this is different from x-step relations, where objects in the view of P would not necessarily be included in the view of O.


== Licenses ==

Licences, in DOMS, have, as their only purpose, to restrict who can view what material. They are only a concern for people using the material in DOMS, not users working with the GUI, or otherwise administrating the contents.

Licences are implemented by using the Fedora XACML engine. When a user authenticates with the Fedora server (or with a server passing authentication tokens to the DOMS), he gets a number of attributes. Each of these attributes name one licence that the user can access material under.

Each data object in DOMS has a POLICY datastream. This datastream is just an URL, referring to a Licence object's LICENCE datastream. This datastream is an XACML stream, that evaluates if the user posses the attribute that specify that the user can use this Licence. If yes, the user is granted access to the original object, otherwise he is denied.


== Audit Trail ==

Each user that will use the GUI will need to login. They will authenticate with some external server, probably the SB LDAP server. The access control is not really a concern for the DOMS system. As such, all GUI users will have equal and full access to the DOMS repository.

Audit trails, however, are a concern. Each change to a datastream in a data object will, per default Fedora behaivour, create a new version of this datastream, marked with the creation time and the username. For this reason the Fedora operations PurgeObject and PurgeDatastream are blocked, as they destroy the audit trail.
Real deletion of information is not possible, but both objects and datastreams can be marked as "deleted", again per standard Fedora behaviour. Any tools working with or on DOMS should respect this flag. The GUI should only concern itself with the most recent version of a datastream.		 Most data models are structured around some realworld concept, like a CD, modelled as a digital object. This object will be described by a content model that is totally collection specific, only extending doms:!ContentModel_DOMS. It will probably have relations to digital objects, like tracks. These will be described by a content model that extends doms:!ContentModel_Audio. Each of these will tracks must then reference a audio preservation file object, or some subtype of this. This is the best practice for constructing data models.

DON'T PANIC

A definition: A datamodel describes the content of a collection. A content model describes the content of a data object. So, a datamodel is a set of content models, that together describe the collection.

The DOMS datamodel describes how the Type system underlying DOMS is realized in Fedora 3.

The entire DOMS datamodel is, in its entirety, a complex system. For proper understanding, the various components have been detailed in seperate documents. Firstly, it consist of a number of extensions to the Fedora system. Secondly, it consist of a number of predefined objects. These objects make use of the extensions to Fedora. Thirdly, it consist of a number of policies for how certain tasks are achived. And fourthly, it consist of a number of API interfaces.

Fedora and DOMS are big on namespaces. To ease writing the documentation, a namespace document, DomsNameSpacesAndSchemas, have been written. All namespaces should be defined there, and all shorthands refer to the namespaces defined therein.

Content Models in general

Fedora provides a repository for digital objects. All objects in the repository can, in principle, be unique, but Fedora provides a way of specifying that an object has a given type. Unfortunately, the type-definitions in Fedora, called Content Models, are rather simplistic by default. We use them as the basis of our type system, with certain enhancements.

For our purposes, there are two kinds of digital objects in Fedora

  • Data objects
  • Content Model objects

The Content Model object, as used in DOMS, describes the compulsory and legal content of an object of this type. It contains the information nessesary to verify if the given object is indeed of this type. For more detail on this, see FedoraOntology and FedoraTypeChecking

A data object can specify the Content Model describing its contents, via a fedora-model:hasModel relation, and in DOMS we require it to be present. A data object will be said to "subcribe" to a Content Model. Content Model inheritance, as specified in FedoraOntology, will be used.

The special Content Model object "doms:ContentModel_DOMS" is the root object. All Content Models must have an "doms-relations:extendsModel" relation to this object, possibly through a number of other Content Models. The complete description of a data object is defined as the set of the descriptions in the Content Model specified with "fedora-model:hasModel" and all Content Models that can be reached from this, by following "doms-relations:extendsModel" relations.

A Content Model can "extend" more than one other Content Model. There is no overriding of Content Models, a subscribing object must be valid in regards to all the Content Models in the inheritance tree.

Content Models have two datastreams in particular that are interesting. These are the ONTOLOGY and DS-COMPOSITE. The Ontology defines the the allowed relations in subscribing objects, and the DS-COMPOSITE defines the required datastreams and any restrictions they must adhere to.

Fedora extensions

Predefined objects

Doms policies

API documentation

Working with the Data model

Doms contains a number of content models. These are meant to serve as the basic buildingblocks for data models for new collections. A datamodel is, of course, not restricted to use only these content models, it can, and should, define it's own. All new content models, should extend doms:ContentModel_DOMS, and all objects that need to reference files outside Fedora should have a content model that derive from doms:ContentModel_File and so on. The content models that provide extra meaning are optional to use, and should at least be extended for the relevant collection.

Most data models are structured around some realworld concept, like a CD, modelled as a digital object. This object will be described by a content model that is totally collection specific, only extending doms:ContentModel_DOMS. It will probably have relations to digital objects, like tracks. These will be described by a content model that extends doms:ContentModel_Audio. Each of these will tracks must then reference a audio preservation file object, or some subtype of this. This is the best practice for constructing data models.

DataModel (last edited 2010-03-17 13:13:00 by localhost)