* Added: ``as_dataframe`` argument to ``audformat.utils.read_csv()`` * Fixed: ``audformat.utils.read_csv()`` now treats float/integer values in ``start``, ``end`` columns as seconds
1.1.2
--------------------------
* Fixed: ``audformat.Database.load()`` when loading databases with a misc table that has an assigned split
1.1.1
--------------------------
* Changed: depend on ``audeer>=2.0.0`` * Fixed: ``pandas`` deprecation warnings
1.1.0
--------------------------
* Added: ``audformat.Database.get()`` method to retrieve labels based on their schemes and independent of the tables in which they are stored * Added: ``aggregate_function`` and ``aggregate_strategy`` arguments to ``audformat.utils.concat()`` to support overlapping values in the objects that should be concatenated * Changed: ``audformat.Column.get(map=...)`` now returns dtype of labels * Changed: ``audformat.Column.get(map=...)`` does no longer raise an error if some of the mapped values are not available when stored in a dictionary as scheme labels * Fixed: avoid deprecation warning by replacing ``pkg_resources`` internally with ``importlib.metadata``
1.0.3
--------------------------
* Fixed: ``audformat.utils.hash()`` for ``pandas>=2.1.0`` * Fixed: remove upper limit of ``pandas`` dependency
1.0.2
--------------------------
* Fixed: require ``pandas<2.1.0`` as ``pandas>=2.1.0`` introduced a bug in calculating the hash of an index * Removed: deprecated ``root`` argument from ``audformat.testing.create_audio_files()``