The workflow in GLAMpipe consists of actions that are done with the data. The actions are called nodes and the workflow is called a pipe.
Nodes are the building blocks of GLAMpipe. The idea of a node is that each node makes one, relatively simple thing with the data or file. By chaining nodes one can make different transformations to the data.
There are four groups of nodes:
Nodes do not overwrite original data. For example, if you use a "split" node for the "author" field, you get a new field called "author_splitted". This makes experimenting easy, since you can not destroy the original data
A user can create a project that contains everything that is used in the project.
A collection is a dataset used in the project. There may be several datasets in a project. They can be used for subsets of data, such as the data for authors or places that are recurrent in the main dataset.