upgini 1.1.167__tar.gz → 1.1.168__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-1.1.167/src/upgini.egg-info → upgini-1.1.168}/PKG-INFO +1 -1
- {upgini-1.1.167 → upgini-1.1.168}/setup.py +1 -1
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/dataset.py +3 -3
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/metrics.py +9 -2
- {upgini-1.1.167 → upgini-1.1.168/src/upgini.egg-info}/PKG-INFO +1 -1
- {upgini-1.1.167 → upgini-1.1.168}/LICENSE +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/README.md +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/pyproject.toml +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/setup.cfg +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/__init__.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/ads.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/ads_management/__init__.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/ads_management/ads_manager.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/data_source/__init__.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/data_source/data_source_publisher.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/errors.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/features_enricher.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/fingerprint.js +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/http.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/mdc/__init__.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/mdc/context.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/metadata.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/normalizer/__init__.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/normalizer/phone_normalizer.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/resource_bundle/__init__.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/resource_bundle/exceptions.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/resource_bundle/strings.properties +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/sampler/__init__.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/sampler/base.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/sampler/random_under_sampler.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/sampler/utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/search_task.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/spinner.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/__init__.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/base_search_key_detector.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/blocked_time_series.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/country_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/custom_loss_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/cv_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/datetime_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/display_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/email_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/features_validator.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/format.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/ip_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/phone_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/postal_code_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/target_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/track_info.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/utils/warning_counter.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini/version_validator.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini.egg-info/SOURCES.txt +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini.egg-info/dependency_links.txt +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini.egg-info/requires.txt +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/src/upgini.egg-info/top_level.txt +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_binary_dataset.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_blocked_time_series.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_categorical_dataset.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_continuous_dataset.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_country_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_custom_loss_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_datetime_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_email_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_etalon_validation.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_features_enricher.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_metrics.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_phone_utils.py +0 -0
- {upgini-1.1.167 → upgini-1.1.168}/tests/test_postal_code_utils.py +0 -0
|
@@ -45,11 +45,11 @@ from upgini.utils.warning_counter import WarningCounter
|
|
|
45
45
|
|
|
46
46
|
class Dataset: # (pd.DataFrame):
|
|
47
47
|
MIN_ROWS_COUNT = 100
|
|
48
|
-
MAX_ROWS =
|
|
48
|
+
MAX_ROWS = 200_000
|
|
49
49
|
FIT_SAMPLE_ROWS = 200_000
|
|
50
50
|
FIT_SAMPLE_THRESHOLD = 200_000
|
|
51
|
-
FIT_SAMPLE_WITH_EVAL_SET_ROWS =
|
|
52
|
-
FIT_SAMPLE_WITH_EVAL_SET_THRESHOLD =
|
|
51
|
+
FIT_SAMPLE_WITH_EVAL_SET_ROWS = 200_000
|
|
52
|
+
FIT_SAMPLE_WITH_EVAL_SET_THRESHOLD = 200_000
|
|
53
53
|
MIN_SAMPLE_THRESHOLD = 20_000
|
|
54
54
|
IMBALANCE_THESHOLD = 0.4
|
|
55
55
|
MIN_TARGET_CLASS_ROWS = 100
|
|
@@ -8,7 +8,15 @@ from catboost import CatBoostClassifier, CatBoostRegressor
|
|
|
8
8
|
from lightgbm import LGBMClassifier, LGBMRegressor
|
|
9
9
|
from numpy import log1p
|
|
10
10
|
from pandas.api.types import is_numeric_dtype
|
|
11
|
-
from sklearn.metrics import
|
|
11
|
+
from sklearn.metrics import check_scoring, get_scorer, make_scorer
|
|
12
|
+
|
|
13
|
+
try:
|
|
14
|
+
from sklearn.metrics import get_scorer_names
|
|
15
|
+
available_scorers = get_scorer_names()
|
|
16
|
+
except ImportError:
|
|
17
|
+
from sklearn.metrics._scorer import SCORERS
|
|
18
|
+
available_scorers = SCORERS
|
|
19
|
+
|
|
12
20
|
from sklearn.metrics._regression import (
|
|
13
21
|
_check_reg_targets,
|
|
14
22
|
check_consistent_length,
|
|
@@ -385,7 +393,6 @@ def _get_scorer(target_type: ModelTaskType, scoring: Union[Callable, str, None])
|
|
|
385
393
|
|
|
386
394
|
multiplier = 1
|
|
387
395
|
if isinstance(scoring, str):
|
|
388
|
-
available_scorers = get_scorer_names()
|
|
389
396
|
metric_name = scoring
|
|
390
397
|
if "mean_squared_log_error" == metric_name or "MSLE" == metric_name or "msle" == metric_name:
|
|
391
398
|
scoring = make_scorer(_ext_mean_squared_log_error, greater_is_better=False)
|
|
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
|
|
File without changes
|
|
File without changes
|