upgini 1.1.307__tar.gz → 1.1.308__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Potentially problematic release.
This version of upgini might be problematic. Click here for more details.
- {upgini-1.1.307 → upgini-1.1.308}/PKG-INFO +1 -1
- upgini-1.1.308/src/upgini/__about__.py +1 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/features_enricher.py +1 -24
- upgini-1.1.307/src/upgini/__about__.py +0 -1
- {upgini-1.1.307 → upgini-1.1.308}/.gitignore +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/LICENSE +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/README.md +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/pyproject.toml +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/__init__.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/ads.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/ads_management/__init__.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/ads_management/ads_manager.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/autofe/__init__.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/autofe/all_operands.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/autofe/binary.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/autofe/date.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/autofe/feature.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/autofe/groupby.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/autofe/operand.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/autofe/unary.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/autofe/vector.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/data_source/__init__.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/data_source/data_source_publisher.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/dataset.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/errors.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/http.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/lazy_import.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/mdc/__init__.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/mdc/context.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/metadata.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/metrics.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/normalizer/__init__.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/normalizer/phone_normalizer.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/resource_bundle/__init__.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/resource_bundle/exceptions.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/resource_bundle/strings.properties +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/resource_bundle/strings_widget.properties +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/sampler/__init__.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/sampler/base.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/sampler/random_under_sampler.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/sampler/utils.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/search_task.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/spinner.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/__init__.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/base_search_key_detector.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/blocked_time_series.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/country_utils.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/custom_loss_utils.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/cv_utils.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/datetime_utils.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/deduplicate_utils.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/display_utils.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/email_utils.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/fallback_progress_bar.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/features_validator.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/format.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/ip_utils.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/phone_utils.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/postal_code_utils.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/progress_bar.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/sklearn_ext.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/target_utils.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/track_info.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/utils/warning_counter.py +0 -0
- {upgini-1.1.307 → upgini-1.1.308}/src/upgini/version_validator.py +0 -0
|
@@ -0,0 +1 @@
|
|
|
1
|
+
__version__ = "1.1.308"
|
|
@@ -90,7 +90,6 @@ from upgini.utils.display_utils import (
|
|
|
90
90
|
from upgini.utils.email_utils import EmailSearchKeyConverter, EmailSearchKeyDetector
|
|
91
91
|
from upgini.utils.features_validator import FeaturesValidator
|
|
92
92
|
from upgini.utils.format import Format
|
|
93
|
-
from upgini.utils.ip_utils import IpToCountrySearchKeyConverter
|
|
94
93
|
from upgini.utils.phone_utils import PhoneSearchKeyDetector
|
|
95
94
|
from upgini.utils.postal_code_utils import PostalCodeSearchKeyDetector
|
|
96
95
|
|
|
@@ -1213,13 +1212,6 @@ class FeaturesEnricher(TransformerMixin):
|
|
|
1213
1212
|
converter = EmailSearchKeyConverter(email_column, hem_column, search_keys, [], self.logger)
|
|
1214
1213
|
extended_X = converter.convert(extended_X)
|
|
1215
1214
|
generated_features.extend(converter.generated_features)
|
|
1216
|
-
if (
|
|
1217
|
-
self.detect_missing_search_keys
|
|
1218
|
-
and list(search_keys.values()) == [SearchKey.DATE]
|
|
1219
|
-
and self.country_code is None
|
|
1220
|
-
):
|
|
1221
|
-
converter = IpToCountrySearchKeyConverter(search_keys, self.logger)
|
|
1222
|
-
extended_X = converter.convert(extended_X)
|
|
1223
1215
|
generated_features = [f for f in generated_features if f in self.fit_generated_features]
|
|
1224
1216
|
|
|
1225
1217
|
return extended_X, search_keys
|
|
@@ -1987,13 +1979,6 @@ class FeaturesEnricher(TransformerMixin):
|
|
|
1987
1979
|
df = converter.convert(df)
|
|
1988
1980
|
generated_features.extend(converter.generated_features)
|
|
1989
1981
|
email_converted_to_hem = converter.email_converted_to_hem
|
|
1990
|
-
if (
|
|
1991
|
-
self.detect_missing_search_keys
|
|
1992
|
-
and list(search_keys.values()) == [SearchKey.DATE]
|
|
1993
|
-
and self.country_code is None
|
|
1994
|
-
):
|
|
1995
|
-
converter = IpToCountrySearchKeyConverter(search_keys, self.logger)
|
|
1996
|
-
df = converter.convert(df)
|
|
1997
1982
|
generated_features = [f for f in generated_features if f in self.fit_generated_features]
|
|
1998
1983
|
|
|
1999
1984
|
meaning_types = {col: key.value for col, key in search_keys.items()}
|
|
@@ -2342,7 +2327,7 @@ class FeaturesEnricher(TransformerMixin):
|
|
|
2342
2327
|
df = self._add_current_date_as_key(df, self.fit_search_keys, self.logger, self.bundle)
|
|
2343
2328
|
|
|
2344
2329
|
# Checks that need validated date
|
|
2345
|
-
validate_dates_distribution(
|
|
2330
|
+
validate_dates_distribution(df, self.fit_search_keys, self.logger, self.bundle, self.warning_counter)
|
|
2346
2331
|
|
|
2347
2332
|
if is_numeric_dtype(df[self.TARGET_NAME]) and has_date:
|
|
2348
2333
|
self._validate_PSI(df.sort_values(by=maybe_date_column))
|
|
@@ -2356,14 +2341,6 @@ class FeaturesEnricher(TransformerMixin):
|
|
|
2356
2341
|
)
|
|
2357
2342
|
df = clean_full_duplicates(df, self.logger, bundle=self.bundle)
|
|
2358
2343
|
|
|
2359
|
-
if (
|
|
2360
|
-
self.detect_missing_search_keys
|
|
2361
|
-
and list(self.fit_search_keys.values()) == [SearchKey.DATE]
|
|
2362
|
-
and self.country_code is None
|
|
2363
|
-
):
|
|
2364
|
-
converter = IpToCountrySearchKeyConverter(self.fit_search_keys, self.logger)
|
|
2365
|
-
df = converter.convert(df)
|
|
2366
|
-
|
|
2367
2344
|
# Explode multiple search keys
|
|
2368
2345
|
non_feature_columns = [self.TARGET_NAME, EVAL_SET_INDEX] + list(self.fit_search_keys.keys())
|
|
2369
2346
|
meaning_types = {
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
__version__ = "1.1.307"
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|