Domain is used to categorize collection of nodes. Generally, nodes of same domain are related by links. But the Graph Portal does not restrict creation of links between nodes of different domains. The sole purpose of domain is organization of nodes.
A node carries the data or information. Node can carry any type of information like HTML fragment, plain text, JSON, XML, CSV data etc. Node editor helps to create and maintain node content. Depending on the type of node content, the application may choose different editors appropriate to the content.
You can mark some of the nodes are as Top Level nodes. Top level nodes will get listed under the domain and serves as the entry point of the graph from the domain. If a node is not a top-level domain, you still can reach node by traversing, search or custom queries.
User can also attach (upload) files to a node. The contents of the attachment is NOT indexed for search. Hence text search will not find keywords inside the attachments.
Deletion of a node will delete all the links to and from the node as well as the attachments of the node.
A link models the relationship between nodes. Graph portal uses directed links (i.e., a link is from a source node to a target node). You can specify bi-directional links using two links between nodes (source --> target and target --> source).
Deletion of a link will NOT delete the nodes attached to it.
A link has a type (aka, Edge Type) and the graph-portal comes with a default type "Has Child". You can create more types as required, through the admin page of the application.