Tamar Grey
Tamar Grey
In #1770, a workaround had to be added for the following pandas bug https://github.com/pandas-dev/pandas/issues/43016. This bug has been fixed and will be released as a part of pandas `1.4.0` release....
In addition to adding primitives that can calculate delayed features in #1730, Featuretools can better support time series problems by investigating existing or potential new primitives that will allow users...
Featuretools could include a `RollingOutlierCount` transform primitive that, like the other rolling primitives, takes in `window_length` and `gap` parameters and allows users to create features that indicate a level of...
closes #1804
There are a number of assumptions we make about the shape of Primitive `input_types` lists: - Its either a list of ColumnSchema objects or a list of lists of ColumnSchema...
- As a user, I wish Featuretools' Feature objects had the ability to store a secondary name for Features. Given that multi-output features exist, this ability to store a second...
When `recalculate_last_time_indexes=True`, it’s possible to end up with not just recalculated last time indexes where they already exist, but with last time indexes calculated on more dataframes than before. For...
While converting Primitives to use Woodwork for their input and return types, there seems to be some common optimizations that can/should be used to define the best input and return...
Currently, `_check_cutoff_time_type` uses a custom `_check_time_type ` util that has slightly different logic for identifying numeric vs datetime columns than Woodwork has. If possible, we should use Woodwork logic for...
In normalize DataFrame, we compile all of the woodwork types that we want to transfer to the new DataFrame in a `transfer_types` dictionary, and then we separate them all out...