Cognite-sdk

Added
- Support for aggregated properties of assets.
- `Asset` and `AssetList` classes now have a `sequences` function which retrieves related sequences.
- Support for partitioned listing of assets and events.

Changed
- `list` and `__call__` methods for assets now support list parameters for `root_ids`, `root_external_ids`.
- Sequences API no longer supports column ids, all relevant functions have been changed to only use external ids.

Fixed
- Fixed a bug in time series pagination where getting 100k dense datapoints would cause a missing id error.
- Sequences retrieve functions fixed to match API change, to single item per retrieve.
- Sequences retrieve/insert functions fixed to match API change to take lists of external ids.

Fixed
- `testing.mock_cognite_client()` so that it still accepts arguments after exiting from mock context.

Fixed
- `testing.mock_cognite_client()` so that the mocked CogniteClient may accept arguments.

Added
- Method `files.download_to_path` for streaming a file to a specific path

Added
- `root_asset_ids` parameter for time series list.

Changed
- Formatted output in jupyter notebooks for `SequenceData`.
- `retrieve_latest` function in theDatapoints API extended to support more than 100 items.
- Log requests at DEBUG level instead of INFO.

Changed
- Disabled automatic handling of cookies on the requests session objects

Fixed
- `to_pandas` method on CogniteResource in the case of objects without metadata