com.google.enterprise.connector.spi

Interface Summary
Interface	Description
AuthenticationIdentity	Interface for the `identity` parameter of the `AuthenticationManager.authenticate` method.
AuthenticationManager	Authentication Manager.
AuthorizationManager	Authorization Manager.
Connector	The root of the SPI for connector instances.
ConnectorFactory	This factory provides a mechanism by which `ConnectorType.validateConfig` may create instances of the connector for the purpose of validation.
ConnectorPersistentStore	Gives connectors access to a local storage area.
ConnectorPersistentStoreAware	If a `Connector` implements this marker interface, then the Connector Manager will call the setter when it instantiates it, supplying a `ConnectorPersistentStore` object, which encapsulates how a connector can persist and use information about documents or any other information.
ConnectorShutdownAware	A `Connector` object may implement this interface if it wishes to be informed when the Connector instance is being shut down or deleted.
ConnectorType	The root of the SPI for connector configuration.
DatabaseResourceBundle	Supplies SQL language syntax variations for the various database implementations.
Document	Interface that represents a document.
DocumentAcceptor	Interface for a DocumentAcceptor - something that takes spi Documents and sends them along on their way.
DocumentList	Interface that represents a list of documents to be traversed.
HasTimeout	Deprecated This Interface has been fully deprecated.
Lister	The `Lister` mechanism might be used as an alternative to the `TraversalManager` as a convenient interface for feeding documents by the connector.
ListerAware	A `Session` should implement this optional interface if it allows `Lister` feed documents.
LocalDatabase	Provides access to the Connector Manager's configured JDBC database which the connector implementer can use for any purpose.
LocalDocumentStore	Deprecated This interface has never been publicly implemented
Property	Interface that represents a property.
Retriever	A `Retriever` provides access to document content, based upon a docid.
RetrieverAware	A `Session` should implement this optional interface if it allows `Retriever` access to document content.
Session	The primary access point to information from the repository.
TraversalContext	A callback interface that the Connector Manager will pass in to a `TraversalManager`, which the traversal manager can the use to call back to get information from the Connector Manager.
TraversalContextAware	A `TraversalManager` should implement this interface if it wants the Connector Manager to supply it with a `TraversalContext` object.
TraversalManager	Interface for implementing query-based traversal.
TraversalSchedule	A traversal schedule for a connector.
TraversalScheduleAware	An optional interface that a `TraversalManager` or `Lister` may implement, which will receive the `TraversalSchedule` for this connector.

Class Summary
Class	Description
AuthenticationResponse	The response from the `AuthenticationManager.authenticate` method.
AuthorizationResponse	The carrier type of the list returned by the `AuthorizationManager.authorizeDocids` method.
ConfigureResponse	The response for most configuration methods, such as `ConnectorType.getConfigForm` and `ConnectorType.validateConfig`.
Principal	Represents a principal for authentication and authorization purposes.
SecureDocument	An implementation of the `Document` interface for use with ACLs.
SimpleAuthenticationIdentity	Simple implementation of the `AuthenticationIdentity` interface.
SimpleConnectorFactory	Simple implementation of the `ConnectorFactory` interface.
SimpleConnectorType	Simple implementation of the `ConnectorType` interface.
SimpleDocument	Simple implementation of the `Document` interface.
SimpleDocumentList	Simple implementation of the `DocumentList` interface.
SimpleProperty	Simple implementation of the `Property` interface.
SimpleTraversalContext	Simple implementation of the `TraversalContext` interface.
SocialCollectionHandler	SocialCollectionHandler initializes a social collection, which is a container in GSA for social documents, like user profile documents.
SocialUserProfileDocument	A user profile document object.
SocialUserProfileDocument.ColleagueData	A colleague object.
SocialUserProfileProperties	Distinguished Properties for user profiles
SpiConstants	Non-instantiable class that holds constants used by the SPI and documents their meanings.
Value	Wrapper class for all data items from a repository.
XmlUtils	Utility class containing methods used to encode and construct XML for the Connector Manager.

Enum Summary
Enum	Description
AuthorizationResponse.Status	Authorization Status codes.
SpiConstants.AclAccess	Enum for the list of possible ACL access values.
SpiConstants.AclInheritanceType	Enum for the list of possible inheritance types.
SpiConstants.AclScope	Enum for the list of possible ACL scope values.
SpiConstants.ActionType	Enum for action types.
SpiConstants.CaseSensitivityType	Enum for the possible case sensitivity rules.
SpiConstants.ContentEncoding	Enum for feed content encoding.
SpiConstants.DatabaseType	This enumeration identifies databases.
SpiConstants.DocumentType	Enum for the list of possible document types.
SpiConstants.FeedType	Enum for the list of possible feed types.
SpiConstants.PrincipalType	Enum for the types of Principals.
SpiConstants.RoleType	Deprecated Roles are ignored by the Google Search Appliance.

Exception Summary
Exception	Description
DocumentAcceptorException	Thrown for general `DocumentAcceptor` errors, such as connectivity problems.
DocumentAccessException	Thrown in cases where the search or traversal user does not have enough privileges to access a document.
DocumentNotFoundException	Thrown for documents that do not exist or are unknown to the connector.
RepositoryDocumentException	Thrown for problems accessing individual documents in a repository.
RepositoryException	Thrown for general repository errors, such as connectivity problems.
RepositoryLoginException	Thrown to indicate unsuccessful login.
SkippedDocumentException	Thrown for documents that should be skipped for non-error reasons.

Package com.google.enterprise.connector.spi Description

Defines the SPI for connector developers who wish to write plug-ins to the Google Search Appliance.

Package Specification

The Connector Manager uses the Spring Framework to instantiate two important interfaces of this package: ConnectorType and Connector. The implementor of this SPI must provide two Spring xml configuration files to notify the Connector Manager of the existence of their implementation and to describe how to instantiate their implementation. These two files must have these names:

connectorType.xml
connectorInstance.xml

The first file, connectorType.xml, should be a complete and independent file that provides a bean definition for a class that implements the ConnectorType interface. The Connector Manager will use the id attribute of that bean as the name of your implementation. It is the implementor's responsibility to make sure that this name is unique in a Search Appliance installation (that is, unique to all Connector Managers in an installation).

When a connector manager starts up, it looks for all classpath resources with this path: config/connectorType.xml. It passes these to Spring which then instantiates them all. Connector developers should include such a file in the jar in which they package their connector.

The second file, connectorInstance.xml, should provide a bean definition for a class that implements the Connector interface. The id attribute for this bean should be the same as the id for your ConnectorType interface. Unlike the connectorType.xml, the file is not normally self-contained. It may refer to external variables in the following syntax:

  <property name="username" value="${username}"/>

These variables will be supplied by Spring from a properties file created by the Connector Manager for each instance of a connector. Spring supplies these values using their PropertyPlaceholderConfigurer functionality. The actual values are expected to come from the Search Appliance administrator, via screens rendered on the Administrator Console.

The primary responsibility of the ConnectorType implementation is to provide the Administrator Console with forms and validation methods for these configuration screens.

Example: suppose you want to write a connector for a 3rd-party CMS named "XyzCms". You would write a complete java implementation of this SPI, using the XyzCms's proprietary API. You would also provide the two xml files, ConnectorType.xml and ConnectorInstance.xml. You might give the defined beans in both XML files the ids "XyzCmsConnector" - or perhaps "XyzCmsConnectorFromMyCompany", to ensure uniqueness.

At a particular customer installation, the Search Appliance administrator would find a machine on which to run an external Connector Manager, would download and install Tomcat, then Connector Manager from the code site, or by using simpler, supported installers, if such exist. He or she would then download and install your connector, according to directions you supply. When the Connector Manager starts up in the new Tomcat instance, it looks for all config files on the class path at "config/connectorType.xml", and it discovers your XyzCmsConnector, perhaps among others. It uses Spring to instantiate your class that implements "ConnectorType" according to that XML file.

Suppose that customer has two different XyzCms repositories, with different names, say "XyzA" and "XyzB". The customer would then use the Search Appliance's Admin Console to create two instances of your connector, perhaps giving the connector instances the same names as the repositories: "XyzA" and "XyzB". In turn, the connector manager would automatically create two properties files, named after the connector instance names, XyzA.properties and XyzB.properties. The connector manager then uses Spring to instantiate two instances of your Connector implementation, using the two properties files.

For details on dependent specifications, see these sites:

Spring Reference Manual, v2.5.6