v1.4.0rc1

• The repository interface to Zenodo has one single upload method upload_file with the parameter metamapper. It
  is a callable which extracts meta information from the actual file to be uploaded. This is especially useful and specifically
  intended for HDF5 files. Unless the value for metamapper is None, the upload_file method will use the built-in
  hdf5 extraction function automatically on HDF5 files.
• Clarify abstraction for HDF5 database interfaces. HDF5DBInterface is the top abstraction from which
  ExtHDF5Interface inherits. ExtHDF5Interface makes use of external databases such as mongoDB.
• fix issue in online documentation: mongomock is used to run the mongodb jupyter notebook in the documentation
• codemeta.json file is updated with author and institution ROR ID

v1.3.0

• important changes:
  • improved and consequent support of RDF/JSON-LD. This means, an HDF5 can be created from a JSON-LD file and vice versa. The JSON-LD file
    contains the structural and contextual metadata of the HDF5 file.
  • namespaces are outsourced to ontolutils
• minor changes:
  • When a file is opened with a filename which does not exist and mode is None, the file will NOT be created. This was
    the case in the past, but this may lead to unwanted empty files.
  • Bugfix namespace creation
  • some method renaming and refactoring
  • accessors are refactored and improved (especially shifted away from xarray and fully integrated in hdf)

v1.0.0

The version v1.0.0 introduced major changes:

• zenodo is not a dependency anymore but is implemented as a new subpackage of the toolbox
• zenodo is part or repository which is designed to provide interfaces to different data repositories (however,
  only zenodo is implemented at the moment)
• the database architecture is changed similarly, such that it has a more logic structure
• both above changes follow a more or less strict inheritance structure from abstract classes defining the interface to
  repositories or databases (databases are meant to be local, like mongoDB, SQL, etc., repositories are online data
  storage, like zenodo, which allows searching for metadata but not within the raw files.)
• python 3.8 until 3.12 inclusive are supported
• IRI as persistent identifier is now supported, which fulfills the F3 requirement of the FAIR principles ("Metadata
  clearly and explicitly include the identifier of the data they describe", https://www.go-fair.org/fair-principles/)
• package renaming and reorganization: conventions is now convention, layout is now a module, new is repository
• usage of IRI (persistent identifier) is now supported

v0.12.2

v0.12.2

• bugfix requirements
• add $exist find()-methods inside of HDF files
• bugfix 0D time data as dimension
• module query functions (find, ...) can guess the filename from objects that have with hasattr(hdf_filename, 'hdf_filename')
• bugfix in zenodo search (did Zenodo change their API?)
• 0D data is written to MongoDB
• new utils like computing filesize
• update to new zenodo_search package due to change in backend at Zenodo.org
• find, find_one and distinct can be called on HDF files
• small bugfixes

v0.11.1

v0.11.1

• bugfixes. See changelog

v0.11.0

v0.11.0

• working with time data is now possible:
  • time data can be created using the high-level method create_time_dataset
  • slicing the above "time datasets" will return xarray data
  • See docs/wrapper/Misc.ipynb
• fixed issue with user-defined & nested standard attributes

v0.10.0

Initial version. A publication is submitted to the journal ing.grid.