Pyramid CMS concepts and architecture

Pyramid and TurboGears

Pyramid CMS is built on top of TurboGears2 framework (TG2 for short). As soon as TG2 is MVC-based, Pyramid is also a MVC application.

Pyramid and other CMSs

Some ideas are inspired by (or borrowed/stolen from) the Drupal.

Publication concept

The main entities are node, comment and attachment. Node is the main entity and each node can have multiple comments and attachmets.

A Node (well, I borrowed the idea from the wonderfull Drupal) is a generic ”publication unit“. Each node has some type (see below) and a unique id, and can be access via a path like /node/13, where 13 is a node's id, and /node/ is a base path for a NodeController (see below).

The Node entity

Each node has a:

1. unique ID
2. title
3. content
4. a reference to a parent node
5. references to child nodes
6. [TODO] references to related nodes
7. references to comments, attachmets, etc.

That is, all nodes can be organized in a tree (or several trees)

Node is extendable via inheritance. In fact, Node is just a base class for more sophisticated entities, like Article, BlogPost or ForumPost. From the database point of view, inheritance used is polymorphic multi-table. That is, all the base preperties (that ones from the base class, Node) are stored in a one table, and the extended attributes of descendants in their own tables. This allows us to access all the nodes, articles, etc. in the same way, perform a search for nodes of any types that have some tags or a subsrting in a title, and, for example, have an Article with a ForumThread or a WikiPage or a Poll attached to it.

Each Node is ”authored” and ”versioned”.