This document explains how Insight imports content from Legend server. It focuses on collaboration between abstract classes, which attempt to hide WebDAV protocol, or file-based storage system.
It also illustrates how Insight Model modification is handled without any lock.

See publications.

See javadoc for this class.

This class defines the API for accessing Insight feature of Publication creation. It completely hides content storage means, and metadata representation.

This package is accessed by CreatePublicationOperation. It relies on the assumption that content (as binary resources) is stored separately from metadata.

See javadoc for this package.

LegendConnector connects to a Legend server, and pulls LegendResources from it.

ImportParameters is a set of parameters. It avoids carrying a great number of parameters in methods.

LegendResources is a set of resources, associating path information, and an input stream.

ImportedResources represent resources taken out from Legend, with an eventual error flag on each.

BLOBStore holds content resources.

StoreFeeder takes care of the details when pulling content from Legend to BLOBStore.

Publication Id creation is a synchronized method of InsightModel. This guarantees that any new PublicationModel instance will have a unique Id. If import fails, the Id is wasted.

Legend resources are enumerated before any further operation.

The case where Legend content structure changes during import is not handled. A later version of Shaman could take advantage of WebDAV locking feature.

Fragments are created from ImportedResources information.
The commit() call at the make the PublicationModel part of the InsightModel.