Pyclarify

Added

- Experimental folder with experimental features. Experimental features might be added to stable or removed.

Changed

- Bumping pydantic version from 1.10 to 2.6.0.
- Removed specific versioning of other dependencies.

Security

- Python 3.7 reached upstream end-of-life on June 27th, 2023, and no longer receives security or bug fixes. Removing Python 3.7 from build order.

Removed

- typing_extensions support is removed, due to Python 3.7 no longer being supported.

Added

- New Evaluate endpoint. A data method for aggregating time-series data and perform evaluate formula expressions. Along with support classes for validation.

python
from pyclarify import ItemAggregation, Calculation

item1 = ItemAggregation(
id="cbpmaq6rpn52969vfl00",
aggregation="max",
alias="i1"
)

item2 = ItemAggregation(
id="cbpmaq6rpn52969vfl0g",
aggregation="avg",
alias="i2"
)

items = [
item1, item2
]


step1 = Calculation(
formula="i1 * i2",
alias="c1"
)

step2 = Calculation(
formula="gapfill(i1) / c1",
alias="c2"
)

calculations = [step1, step2]

r = client.evaluate(
items=items,
calculations=calculations,
gte="2023-10-01T00:00:00Z",
lt="2023-10-23T00:00:00Z",
include=["items"],
rollup="PT10M"
)

>>> r.result.data.to_pandas()
... c1 c2 i1 i2
... 2023-10-20 10:20:00+00:00 21.333333 0.281250 6.0 3.555556
... 2023-10-20 10:30:00+00:00 42.428571 0.212121 9.0 4.714286
... 2023-10-20 10:40:00+00:00 36.363636 0.220000 8.0 4.545455


Changed

- API version changed from 1.1beta2 to 1.1.
- Changes to internal models to comply with new models in the API.
- Added `origin`, `timeZone`, and `firstDayOfWeek` parameters to `pyclarify.Client.data_frame` method, allowing higher control of rollup options.

Fixed

- Typo in CONTRIBUTION.md.

Fixed

- Still experienced issues. Forced SDK to use pydantic.V1.

Fixed

- Fixed deprecated `Pydantic` dependency to the `pydantic.fields.Optional` resource. Changed all occurances to `typing.Optional`.

Added

- Added possibility to directly insert pd.DataFrame, pd.Series and dictionaries into Clarify. NB! Infers timestamp column and only supports when there only exists one timestamp column.
- Added possibility to directly insert dictionaries to `pyclarify.client.data_frame`, `pyclarify.client.select_signals`, and `pyclarify.client.select_items`.
- Added default timezone for general timestamp such as `datetime.today()`. This also applies to timestamps without explicit timezone.

Fixed

- Fixed bug where the response was not JSON encoded causing error initialization to fail. Added more tests to check for such issues.
- Fixed bug where Filter did not allow dictionary without comparison as an input argument for the filter.
- Fixed included list containing multiple instances of same object.