Openeo

Added

- Automatically use `load_url` when providing a URL as geometries to `DataCube.aggregate_spatial()`, `DataCube.mask_polygon()`, etc. ([104](https://github.com/Open-EO/openeo-python-client/issues/104), [#457](https://github.com/Open-EO/openeo-python-client/issues/457))
- Allow specifying `limit` when listing batch jobs with `Connection.list_jobs()` ([677](https://github.com/Open-EO/openeo-python-client/issues/677))
- Add `additional` and `job_options` arguments to `Connection.download()`, `Datacube.download()` and related ([681](https://github.com/Open-EO/openeo-python-client/issues/681))

Changed

- `MultiBackendJobManager`: costs has been added as a column in tracking databases ([[588](https://github.com/Open-EO/openeo-python-client/issues/588)])
- When passing a path/string as `geometry` to `DataCube.aggregate_spatial()`, `DataCube.mask_polygon()`, etc.:
this is not translated automatically anymore to deprecated, non-standard `read_vector` usage.
Instead, if it is a local GeoJSON file, the GeoJSON data will be loaded directly client-side.
([104](https://github.com/Open-EO/openeo-python-client/issues/104), [#457](https://github.com/Open-EO/openeo-python-client/issues/457))
- Move `read()` method from general `JobDatabaseInterface` to more specific `FullDataFrameJobDatabase` ([680](https://github.com/Open-EO/openeo-python-client/issues/680))
- Align `additional` and `job_options` arguments in `Connection.create_job()`, `DataCube.create_job()` and related.
Also, follow official spec more closely. ([683](https://github.com/Open-EO/openeo-python-client/issues/683), [Open-EO/openeo-api#276](https://github.com/Open-EO/openeo-api/issues/276))

Fixed

- `load_stac`: use fallback temporal dimension when no "cube:dimensions" in STAC Collection ([666](https://github.com/Open-EO/openeo-python-client/issues/666))
- Fix usage of `Parameter.spatial_extent()` with `load_collection` and `filter_bbox` ([676](https://github.com/Open-EO/openeo-python-client/issues/676))

Added

- Added `MultiResult` helper class to build process graphs with multiple result nodes ([391](https://github.com/Open-EO/openeo-python-client/issues/391))

Fixed

- `MultiBackendJobManager`: Fix issue with duplicate job starting across multiple backends ([654](https://github.com/Open-EO/openeo-python-client/pull/654))
- `MultiBackendJobManager`: Fix encoding issue of job metadata in `on_job_done` ([657](https://github.com/Open-EO/openeo-python-client/issues/657))
- `MultiBackendJobManager`: Avoid `SettingWithCopyWarning` ([641](https://github.com/Open-EO/openeo-python-client/issues/641))
- Avoid creating empty file if asset download request failed.
- `MultiBackendJobManager`: avoid dtype loading mistakes in `CsvJobDatabase` on empty columns ([656](https://github.com/Open-EO/openeo-python-client/issues/656))
- `MultiBackendJobManager`: restore logging of job status histogram during `run_jobs` ([655](https://github.com/Open-EO/openeo-python-client/issues/655))

Removed

- Drop support for Python 3.7 ([578](https://github.com/Open-EO/openeo-python-client/issues/578))

Fixed

- Fixed broken support for `title` and `description` job properties in `execute_batch()` ([652](https://github.com/Open-EO/openeo-python-client/issues/652))

Added

- Added `DataCube.load_stac()` to also support creating a `load_stac` based cube without a connection ([638](https://github.com/Open-EO/openeo-python-client/issues/638))
- `MultiBackendJobManager`: Added `initialize_from_df(df)` (to `CsvJobDatabase` and `ParquetJobDatabase`) to initialize (and persist) the job database from a given DataFrame.
Also added `create_job_db()` factory to easily create a job database from a given dataframe and its type guessed from filename extension.
([635](https://github.com/Open-EO/openeo-python-client/issues/635))
- `MultiBackendJobManager.run_jobs()` now returns a dictionary with counters/stats about various events during the full run of the job manager ([645](https://github.com/Open-EO/openeo-python-client/issues/645))
- Added (experimental) `ProcessBasedJobCreator` to be used as `start_job` callable with `MultiBackendJobManager` to create multiple jobs from a single parameterized process (e.g. a UDP or remote process definition) ([604](https://github.com/Open-EO/openeo-python-client/issues/604))

Fixed

- When using `DataCube.load_collection()` without a connection, it is not necessary anymore to also explicitly set `fetch_metadata=False` ([638](https://github.com/Open-EO/openeo-python-client/issues/638))

Added

- `load_stac`/`metadata_from_stac`: add support for extracting actual temporal dimension metadata ([567](https://github.com/Open-EO/openeo-python-client/issues/567))
- `MultiBackendJobManager`: add `cancel_running_job_after` option to automatically cancel jobs that are running for too long ([590](https://github.com/Open-EO/openeo-python-client/issues/590))
- Added `openeo.api.process.Parameter` helper to easily create a "spatial_extent" UDP parameter
- Wrap OIDC token request failure in more descriptive `OidcException` (related to [624](https://github.com/Open-EO/openeo-python-client/issues/624))
- Added `auto_add_save_result` option (on by default) to disable automatic addition of `save_result` node on `download`/`create_job`/`execute_batch` ([513](https://github.com/Open-EO/openeo-python-client/issues/513))
- Add support for `apply_vectorcube` UDF signature in `run_udf_code` ([Open-EO/openeo-geopyspark-driver881](https://github.com/Open-EO/openeo-geopyspark-driver/issues/811))
- `MultiBackendJobManager`: add API to the update loop in a separate thread, allowing controlled interruption.

Changed

- `MultiBackendJobManager`: changed job metadata storage API, to enable working with large databases
- `DataCube.apply_polygon()`: rename `polygons` argument to `geometries`, but keep support for legacy `polygons` for now ([592](https://github.com/Open-EO/openeo-python-client/issues/592), [#511](https://github.com/Open-EO/openeo-processes/issues/511))
- Disallow ambiguous single string argument in `DataCube.filter_temporal()` ([628](https://github.com/Open-EO/openeo-python-client/issues/628))
- Automatic adding of `save_result` from `download()` or `create_job()`:
inspect whole process graph for pre-existing `save_result` nodes
(related to [623](https://github.com/Open-EO/openeo-python-client/issues/623), [#401](https://github.com/Open-EO/openeo-python-client/issues/401), [#583](https://github.com/Open-EO/openeo-python-client/issues/583))
- Disallow ambiguity of combining explicit `save_result` nodes
and implicit `save_result` addition from `download()`/`create_job()` calls with `format`
(related to [623](https://github.com/Open-EO/openeo-python-client/issues/623), [#401](https://github.com/Open-EO/openeo-python-client/issues/401), [#583](https://github.com/Open-EO/openeo-python-client/issues/583))

Fixed

- `apply_dimension` with a `target_dimension` argument was not correctly adjusting datacube metadata on the client side, causing a mismatch.
- Preserve non-spatial dimension metadata in `aggregate_spatial` ([612](https://github.com/Open-EO/openeo-python-client/issues/612))

Added

- Add experimental `openeo.testing.results` subpackage with reusable test utilities for comparing batch job results with reference data
- `MultiBackendJobManager`: add initial support for storing job metadata in Parquet file (instead of CSV) ([571](https://github.com/Open-EO/openeo-python-client/issues/571))
- Add `Connection.authenticate_oidc_access_token()` to set up authorization headers with an access token that is obtained "out-of-band" ([598](https://github.com/Open-EO/openeo-python-client/issues/598))
- Add `JobDatabaseInterface` to allow custom job metadata storage with `MultiBackendJobManager` ([571](https://github.com/Open-EO/openeo-python-client/issues/571))