Skip to content

Specifications for pre-defined data structures provided by HDMF.

License

Notifications You must be signed in to change notification settings

hdmf-dev/hdmf-common-schema

Repository files navigation

hdmf-common

Documentation of the HDMF Common data format specification is available at https://hdmf-common-schema.readthedocs.io

Citing HDMF Common Schema

  • RRID: HDMF Common Schema, RRID:SCR_021342

Description

The HDMF Common specification defines a collection of common, reusable data structures that build the foundation for the modeling of advanced data formats, e.g., the Neurodata Without Borders (NWB) neurophysiology data standard. The HDMF Common schema is integrated with HDMF, which provides advanced APIs for reading, writing, and using HDMF-common data types.

The HDMF-common schema provides the following data structures:

  • DynamicTable: A column-based table data structure that supports ragged columns and one-to-one and one-to-many relationships.
  • VectorData: A data structure for representing a column of a DynamicTable.
  • VectorIndex: A data structure for indexing a VectorData. This is used to store one-to-many relationships.
  • ElementIdentifiers: A 1D array for storing primary identifiers for elements of a table.
  • DynamicTableRegion: A data structure for linking to a row or set of rows of a DynamicTable.
  • AlignedDynamicTable: A DynamicTable that supports storing a collection of sub-tables.
  • CSRMatrix: A compressed sparse row matrix.

The schema also provides the following base data structures:

  • Data: An abstract data type for a dataset.
  • Container: An abstract data type for a generic container storing collections of data and metadata, i.e., a group.
  • SimpleMultiContainer: A simple container that holds multiple Data and Container types.

Finally, HDMF-common contains experimental data structures. Prior to adding a new data type to the HDMF-common specification, new data structures are added to the HDMF-experimental to enable users to experiment with these data structures. Because these data structures are experimental, they are not guaranteed to maintain backward compatibility, and may never make it into HDMF-common.

Current experimental data types are:

  • ExternalResources: A data structure that contains row-based compound-data tables for storing ontology information and other external resource references.
  • EnumData: A data structure for representing a column where the data come from a fixed set of elements.

Generate documentation

pip install -r requirements-doc.txt
cd docs
make fulldoc