An open source python library for automated feature engineering
BSD-3-CLAUSE License
Bot releases are hidden (Show)
Published by tamargrey almost 3 years ago
Development release for testing purposes
Published by rwedge almost 3 years ago
base_of_exclude
attribute on primitive instead feature class (#1749)get_unused_primitives
only recognizes lowercase primitive strings (#1733)Thanks to the following people for contributing to this release:
@bchen1116, @gsheni, @HenryRocha, @jeff-hernandez, @ridicolos, @rwedge
Published by rwedge almost 3 years ago
Development release for testing purposes
Published by thehomebrewnerd about 3 years ago
v1.0.0 Oct 12, 2021
EntitySet.__deepcopy__
that retains Woodwork typing information (#1465)EntitySet.__getstate__
and EntitySet.__setstate__
to preserve typing when pickling (#1581)DFSTransformer
Documentation for Featuretools 1.0 (#1605)calculate_feature_matrix
time type check and encode_features
for synthesis tests (#1580)Equal
and NotEqual
primitives (#1640)EntitySet.add_relationship
that caused foreign_key
tag to be lost (#1675)add_interesting_values
from Entity
(#1269)set_secondary_time_index
method from Entity
to EntitySet
(#1280)Entity.update_data
with EntitySet.update_dataframe
(#1398)EntitySet
(#1400)Entity
objects in EntitySet
with Woodwork dataframes (#1405)EntitySet.plot
to work with Woodwork dataframes (#1468)last_time_index
to be a column on the DataFrame (#1456)EntitySet.query_by_values
to work with Woodwork dataframes (#1467)list_variable_types
with list_logical_types
(#1477)EntitySet.concat
to work with Woodwork DataFrames (#1490)remove_highly_correlated_features
if necessary (#1618)CumCount
primitive (#1651)entity_from_dataframe
with add_dataframe
(#1504)generate_description
(#1403)LogicalType.transform
and LogicalType instances (#1451)ColumnSchema
for primitive input and return types (#1411)Entity
and Variable
classes (#1501)make_index
functionality to EntitySet (#1507)ColumnSchema
in DFS primitive matching (#1523)add_interesting_values
(#1550)calculate_feature_matrix
to use Woodwork (#1533)nlp-primitives
requirement for Featuretools 1.0 (#1609)Entity
and Variable
in code (#1612)_dataframes_equal
and camel_to_snake
functions (#1683)encode_features
(#1694)encode_features
if not inplace, some space cost (#1699)using_entitysets
page to use Woodwork (#1532), :pr:
1708, :pr:
1713, :pr:
1716)contributing.md
to improve instructions for external contributors (#1723)1677
and :pr:1679
. The related bug in pandas still exists. (#1731)EntitySet
tests (#1548)Thanks to the following people for contributing to this release:
@bchen1116, @gsheni , @HenryRocha, @jeff-hernandez, @rwedge, @tamargrey, @thehomebrewnerd, @VaishnaviNandakumar
Published by thehomebrewnerd about 3 years ago
Development release for testing purposes
Published by rwedge about 3 years ago
Release candidate for version 1.0
For additional documentation, check out the 1.0 transition guide
Adding Interesting Values
To add interesting values for a single entity, call EntitySet.add_interesting_values
passing the
id of the entity for which interesting values should be added.
>>> es.add_interesting_values(entity_id='log')
Setting a Secondary Time Index
To set a secondary time index for a specific entity, call EntitySet.set_secondary_time_index
passing
Entity for which to set the secondary time index along with the dictionary mapping the secondary time
index column to the for which the secondary time index applies.
>>> customers_secondary_time_index = {'cancel_date': ['cancel_reason']}
>>> es.set_secondary_time_index(es['customers'], customers_secondary_time_index)
Creating a Relationship and Adding to an EntitySet
Relationships are now created by passing parameters identifying the entityset along with four string values
specifying the parent dataframe, parent column, child dataframe and child column. Specifying parameter names
is optional.
>>> new_relationship = Relationship(
... entityset=es,
... parent_dataframe_name='customers',
... parent_column_name='id',
... child_dataframe_name='sessions',
... child_column_name='customer_id'
... )
Relationships can now be added to EntitySets in one of two ways. The first approach is to pass in
name values for the parent dataframe, parent column, child dataframe and child column. Specifying
parameter names is optional with this approach.
>>> es.add_relationship(
... parent_dataframe_name='customers',
... parent_column_name='id',
... child_dataframe_name='sessions',
... child_column_name='customer_id'
... )
Relationships can also be added by passing in a previously created Relationship
object. When using
this approach the relationship
parameter name must be included.
>>> es.add_relationship(relationship=new_relationship)
Replace DataFrame
To replace a dataframe in an EntitySet with a new dataframe, call EntitySet.replace_dataframe
and pass in the name of the dataframe to replace along with the new data.
>>> es.replace_dataframe(dataframe_name='log', df=df)
List Logical Types and Semantic Tags
Logical types and semantic tags have replaced variable types to parse and interpret columns. You can list all the available logical types by calling featuretools.list_logical_types
.
>>> ft.list_logical_types()
You can list all the available semantic tags by calling featuretools.list_semantic_tags
.
>>> ft.list_semantic_tags()
Entity.add_interesting_values
has been removed. To add interesting values for a singleEntitySet.add_interesting_values
and pass the name of the dataframe fordataframe_name
parameter (#1405, #1370).Entity.set_secondary_time_index
has been removed and replaced by EntitySet.set_secondary_time_index
dataframe_name
parameter to specify the dataframe on which to set the secondary time index (#1405, #1370).Relationship
initialization has been updated to accept four name values for the parent dataframe,Variable
objects (#1405, #1370).EntitySet.add_relationship
has been updated to accept dataframe and column name values or aRelationship
object. Adding a relationship from a Relationship
object now requires passingEntity.update_data
has been removed. To update the dataframe, call EntitySet.replace_dataframe
and use the dataframe_name
parameter (#1630, #1522).EntitySet
is no longer stored in Entity
objects. Instead, dataframesEntitySet.__init__
has replaced the variable_types
elementlogical_types
and semantic_tags
dictionaries (#1405).EntitySet.entity_from_dataframe
no longer exists. To add new tables to an entityset, useEntitySet.add_dataframe
(#1405).EntitySet.normalize_entity
has been renamed to EntitySet.normalize_dataframe
(#1405).EntitySet.add_relationship
when the dtypes of parent and child columnsEntitySet.add_dataframe
, the first column will only be used as index if'numeric'
semantic tag string, and a datetime time typeDatetime
logical type (#1405).last_time_index
, secondary_time_index
, and interesting_values
are no longer attributeslist_variable_types
will be removed in a future release and replaced by list_logical_types
.list_variable_types
will return the same output as list_logical_types
(#1447).EntitySet.__deepcopy__
that retains Woodwork typing information (#1465)EntitySet.__getstate__
and EntitySet.__setstate__
to preserve typing when pickling (#1581)DFSTransformer
Documentation for Featuretools 1.0 (#1605)calculate_feature_matrix
time type check and encode_features
for synthesis tests (#1580)Equal
and NotEqual
primitives (#1640)EntitySet.add_relationship
that caused foreign_key
tag to be lost (#1675)add_interesting_values
from Entity
(#1269)set_secondary_time_index
method from Entity
to EntitySet
(#1280)Entity.update_data
with EntitySet.update_dataframe
(#1398)EntitySet
(#1400)Entity
objects in EntitySet
with Woodwork dataframes (#1405)EntitySet.plot
to work with Woodwork dataframes (#1468)last_time_index
to be a column on the DataFrame (#1456)EntitySet.query_by_values
to work with Woodwork dataframes (#1467)list_variable_types
with list_logical_types
(#1477)EntitySet.concat
to work with Woodwork DataFrames (#1490)remove_highly_correlated_features
if necessary (#1618)CumCount
primitive (#1651)entity_from_dataframe
with add_dataframe
(#1504)generate_description
(#1403)LogicalType.transform
and LogicalType instances (#1451)ColumnSchema
for primitive input and return tyes (#1411)Entity
and Variable
classes (#1501)make_index
functionality to EntitySet (#1507)ColumnSchema
in DFS primitive matching (#1523)add_interesting_values
(#1550)calculate_feature_matrix
to use Woodwork (#1533)nlp-primitives
requirement for Featuretools 1.0 (#1609)Entity
and Variable
in code (#1612)_dataframes_equal
and camel_to_snake
functions (#1683)encode_features
(#1694)encode_features
if not inplace, some space cost (#1699)using_entitysets
page to use Woodwork (#1532)EntitySet
tests (#1548)Thanks to the following people for contributing to this release:
@gsheni, @jeff-hernandez, @rwedge, @tamargrey, @thehomebrewnerd
Published by thehomebrewnerd about 3 years ago
Development release for testing purposes
Published by thehomebrewnerd about 3 years ago
v0.27.1 Sep 2, 2021
Thanks to the following people for contributing to this release:
@thehomebrewnerd
Published by jeff-hernandez about 3 years ago
Changes
featuretools_primitives
entrypoint throws an exception (#1662)RuntimeError
when two primitives with the same name are encountered during featuretools_primitives
entrypoint handling (#1662)featuretools_primitives
entrypoint loader from loading non-class objects as well as the AggregationPrimitive
and TransformPrimitive
base classes (#1662)Testing Changes
Thanks to the following people for contributing to this release:
@davesque, @gsheni, @jeff-hernandez, @rwedge
Published by jeff-hernandez about 3 years ago
Development release for testing purposes
Published by simha104 about 3 years ago
Thanks to the following people for contributing to this release:
@gsheni, @simha104
Published by simha104 about 3 years ago
Development release for testing purposes
Published by davesque about 3 years ago
name
attribute for EmailAddressToDomain
primitive (#1543)Thanks to the following people for contributing to this release:
@davesque, @rwedge
Published by davesque about 3 years ago
Development release for testing purposes
Published by tuethan1999 over 3 years ago
replace_inf_values
utility function for replacing inf
values in a feature matrix (#1505)include_entities
correctly overrides exclude_entities
in primitive_options
(#1518)Published by tuethan1999 over 3 years ago
Development release for testing purposes
Published by tuethan1999 over 3 years ago
Development release for testing purposes
Published by rwedge over 3 years ago
get_valid_primitives
function (#1462)EntitySet.dataframe_type
attribute (#1473)Thanks to the following people for contributing to this release: @gsheni, @rwedge
Published by rwedge over 3 years ago
Development release for testing purposes
Published by jeff-hernandez over 3 years ago
Thanks to the following people for contributing to this release:
@gsheni, @jeff-hernandez, @rwedge, @thehomebrewnerd