Skip to content

Concepts

This page defines the concepts and terms used throughout madbot.

A workspace is where a project lives in madbot. You create a workspace, invite collaborators and give each of them a role. Workspaces keep projects and their contributors apart: each workspace is independent from the others.

Each workspace has a dashboard that gives an overview of its data, the recent activity and the state of its submissions.

Nodes organise the project inside a workspace. They form trees that represent the parts of a research project: a campaign and its sites, a cohort and its visits, a collection and its boxes.

Each node holds:

  • a title and a description, in rich text;
  • data, linked where they are stored;
  • metadata, which describe the content of the node;
  • samples, the objects the project studies;
  • child nodes.

Metadata flow down the tree: a value set on a node applies to its child nodes. You can override an inherited value at any level, while the rest of the tree keeps the shared value.

Workspaces and nodes have members, each with a role. A workspace or a node can have several owners, so that its management is shared. The roles reference lists what each role can do.

Plugins connect madbot to other systems. There are three kinds:

  • data plugins give access to the files and folders of a source;
  • submission plugins send data and metadata to a repository;
  • metadata field definition plugins add metadata fields from a standard to the madbot referential.

The plugins reference lists the available plugins.

A datalink links data to a madbot node without copying the data into madbot. Data plugins give access to the data and create the datalinks.

  • No copy: the data stay where they are.
  • Access on demand: madbot downloads the data only when a submission needs them.
  • One plugin per kind of source: each data plugin supports one kind of storage.

A data association groups several data together and can carry its own metadata. For example:

  • the two files of paired-end reads;
  • processed data and the raw data they come from;
  • an analysis script, its input data and its results.

Samples are the objects a project studies: a soil core, a material specimen, a cohort participant, an archive document, a biological specimen. They belong to nodes and carry their own metadata.

Metadata describe research data and give them context. They can be attached to:

  • nodes;
  • samples;
  • data;
  • data associations.

Metadata are inherited:

  • down the node tree: a value set on a node applies to its child nodes;
  • from samples to data: data inherit the metadata of the samples they are associated with;
  • with overrides: an inherited value can be changed at any level.

A submission sends data and metadata to an external repository. Submission plugins gather the metadata, organise the files, check them and send them to the repository. This spares manual copying, keeps metadata consistent across platforms and records what was sent.

  1. Collect the metadata of the submitted data.
  2. Organise the data as the repository expects them.
  3. Validate the metadata and the data before sending.
  4. Transfer them to the repository.

Metadata objects hold the metadata of a submission in units that a repository understands. They mirror the containers of the target repository: for example a study, a sample or an experiment.