upgini 1.2.29a5__tar.gz → 1.2.29a7__tar.gz

{upgini-1.2.29a5 → upgini-1.2.29a7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: upgini
-Version: 1.2.29a5
+Version: 1.2.29a7
 Summary: Intelligent data search & enrichment for Machine Learning
 Project-URL: Bug Reports, https://github.com/upgini/upgini/issues
 Project-URL: Homepage, https://upgini.com/

upgini-1.2.29a7/src/upgini/__about__.py

@@ -0,0 +1 @@
+__version__ = "1.2.29a7"

{upgini-1.2.29a5 → upgini-1.2.29a7}/src/upgini/features_enricher.py

@@ -2,6 +2,7 @@ import dataclasses
 import datetime
 import gc
 import hashlib
+import itertools
 import logging
 import numbers
 import os
@@ -53,7 +54,6 @@ from upgini.metadata import (
     SYSTEM_RECORD_ID,
     TARGET,
     CVType,
-    FeaturesMetadataV2,
     FileColumnMeaningType,
     ModelTaskType,
     RuntimeParameters,
@@ -159,6 +159,10 @@ class FeaturesEnricher(TransformerMixin):
 
     shared_datasets: list of str, optional (default=None)
         List of private shared dataset ids for custom search
+
+    select_features: bool, optional (default=False)
+        If True, return only selected features both from input and data sources.
+        Otherwise, return all features from input and only selected features from data sources.
     """
 
     TARGET_NAME = "target"
@@ -279,7 +283,7 @@ class FeaturesEnricher(TransformerMixin):
         self._relevant_data_sources_wo_links: pd.DataFrame = self.EMPTY_DATA_SOURCES
         self.metrics: Optional[pd.DataFrame] = None
         self.feature_names_ = []
-        self.client_feature_names_ = []
+        self.dropped_client_feature_names_ = []
         self.feature_importances_ = []
         self.search_id = search_id
         self.select_features = select_features
@@ -1443,7 +1447,11 @@ class FeaturesEnricher(TransformerMixin):
         client_features = [
             c
             for c in X_sampled.columns.to_list()
-            if (not self.select_features or c in self.feature_names_)
+            if (
+                not self.select_features
+                or c in self.feature_names_
+                or (self.fit_columns_renaming is not None and self.fit_columns_renaming.get(c) in self.feature_names_)
+            )
             and c
             not in (
                 excluding_search_keys
@@ -1660,7 +1668,10 @@ class FeaturesEnricher(TransformerMixin):
         generated_features = []
         if date_column is not None:
             converter = DateTimeSearchKeyConverter(date_column, self.date_format, self.logger, self.bundle)
-            df = converter.convert(df, keep_time=True)
+            # Leave original date column values
+            df_with_date_features = converter.convert(df, keep_time=True)
+            df_with_date_features[date_column] = df[date_column]
+            df = df_with_date_features
             generated_features = converter.generated_features
 
         email_columns = SearchKey.find_all_keys(search_keys, SearchKey.EMAIL)
@@ -1669,9 +1680,10 @@ class FeaturesEnricher(TransformerMixin):
             df = generator.generate(df)
             generated_features.extend(generator.generated_features)
 
-        normalizer = Normalizer(self.bundle, self.logger)
-        df, search_keys, generated_features = normalizer.normalize(df, search_keys, generated_features)
-        columns_renaming = normalizer.columns_renaming
+        # normalizer = Normalizer(self.bundle, self.logger)
+        # df, search_keys, generated_features = normalizer.normalize(df, search_keys, generated_features)
+        # columns_renaming = normalizer.columns_renaming
+        columns_renaming = {c: c for c in df.columns}
 
         df, _ = clean_full_duplicates(df, logger=self.logger, bundle=self.bundle)
 
@@ -2071,7 +2083,7 @@ class FeaturesEnricher(TransformerMixin):
             is_demo_dataset = hash_input(validated_X) in DEMO_DATASET_HASHES
 
             columns_to_drop = [
-                c for c in validated_X.columns if c in self.feature_names_ and c not in self.client_feature_names_
+                c for c in validated_X.columns if c in self.feature_names_ and c in self.dropped_client_feature_names_
             ]
             if len(columns_to_drop) > 0:
                 msg = self.bundle.get("x_contains_enriching_columns").format(columns_to_drop)
@@ -2101,7 +2113,7 @@ class FeaturesEnricher(TransformerMixin):
             date_column = SearchKey.find_key(search_keys, [SearchKey.DATE, SearchKey.DATETIME])
             if date_column is not None:
                 converter = DateTimeSearchKeyConverter(date_column, self.date_format, self.logger, bundle=self.bundle)
-                df = converter.convert(df)
+                df = converter.convert(df, keep_time=True)
                 self.logger.info(f"Date column after convertion: {df[date_column]}")
                 generated_features.extend(converter.generated_features)
             else:
@@ -2196,11 +2208,12 @@ class FeaturesEnricher(TransformerMixin):
 
             if add_fit_system_record_id:
                 df = self.__add_fit_system_record_id(df, search_keys, SYSTEM_RECORD_ID)
-                if DateTimeSearchKeyConverter.DATETIME_COL in df.columns:
-                    df = df.drop(columns=DateTimeSearchKeyConverter.DATETIME_COL)
                 df = df.rename(columns={SYSTEM_RECORD_ID: SORT_ID})
                 features_not_to_pass.append(SORT_ID)
 
+            if DateTimeSearchKeyConverter.DATETIME_COL in df.columns:
+                df = df.drop(columns=DateTimeSearchKeyConverter.DATETIME_COL)
+
             # search keys might be changed after explode
             columns_for_system_record_id = sorted(list(search_keys.keys()) + features_for_transform)
             df[SYSTEM_RECORD_ID] = pd.util.hash_pandas_object(df[columns_for_system_record_id], index=False).astype(
@@ -2219,7 +2232,7 @@ class FeaturesEnricher(TransformerMixin):
 
             combined_search_keys = combine_search_keys(search_keys.keys())
 
-            df_without_features = df.drop(columns=features_not_to_pass)
+            df_without_features = df.drop(columns=features_not_to_pass, errors="ignore")
 
             df_without_features, full_duplicates_warning = clean_full_duplicates(
                 df_without_features, self.logger, bundle=self.bundle
@@ -2328,11 +2341,15 @@ class FeaturesEnricher(TransformerMixin):
             else:
                 result = enrich()
 
+            selecting_columns = [
+                c
+                for c in itertools.chain(validated_X.columns.tolist(), generated_features)
+                if c not in self.dropped_client_feature_names_
+            ]
             filtered_columns = self.__filtered_enriched_features(importance_threshold, max_features)
-            existing_filtered_columns = [
+            selecting_columns.extend(
                 c for c in filtered_columns if c in result.columns and c not in validated_X.columns
-            ]
-            selecting_columns = validated_X.columns.tolist() + generated_features + existing_filtered_columns
+            )
             if add_fit_system_record_id:
                 selecting_columns.append(SORT_ID)
 
@@ -3510,7 +3527,7 @@ class FeaturesEnricher(TransformerMixin):
         features_df = self._search_task.get_all_initial_raw_features(trace_id, metrics_calculation=True)
 
         self.feature_names_ = []
-        self.client_feature_names_ = []
+        self.dropped_client_feature_names_ = []
         self.feature_importances_ = []
         features_info = []
         features_info_without_links = []
@@ -3520,19 +3537,23 @@ class FeaturesEnricher(TransformerMixin):
         for feature_meta in features_meta:
             if feature_meta.name in original_names_dict.keys():
                 feature_meta.name = original_names_dict[feature_meta.name]
+
+            is_client_feature = feature_meta.name in x_columns
+
+            if feature_meta.shap_value == 0.0:
+                if self.select_features:
+                    self.dropped_client_feature_names_.append(feature_meta.name)
+                continue
+
             # Use only important features
             if (
-                (feature_meta.shap_value == 0.0)
-                or (feature_meta.name in self.fit_generated_features)
-                or (feature_meta.name == COUNTRY)
+                feature_meta.name in self.fit_generated_features
+                or feature_meta.name == COUNTRY
+                # In select_features mode we select also from etalon features and need to show them
+                or (not self.select_features and is_client_feature)
             ):
                 continue
 
-            is_client_feature = feature_meta.name in x_columns
-            # In select_features mode we select also from etalon features and need to show them
-            if not self.select_features and is_client_feature:
-                continue
-
             self.feature_names_.append(feature_meta.name)
             self.feature_importances_.append(_round_shap_value(feature_meta.shap_value))
 

{upgini-1.2.29a5 → upgini-1.2.29a7}/src/upgini/http.py

@@ -882,7 +882,7 @@ class _RestClient:
         if content_type:
             headers[_RestClient.CONTENT_TYPE_HEADER_NAME] = content_type
         if trace_id:
-            headers[_RestClient.TRACE_ID_HEADER_NAME] = trace_id
+            headers[_RestClient.TRACE_ID_HEADER_NAME] = str(trace_id)
         for header_key, header_value in additional_headers.items():
             headers[header_key] = header_value
         return headers

{upgini-1.2.29a5 → upgini-1.2.29a7}/src/upgini/utils/datetime_utils.py

@@ -114,10 +114,12 @@ class DateTimeSearchKeyConverter:
             period_suffix = f"_{period}" if column != "day_in_quarter" else ""
             sin_feature = f"datetime_{column}_sin{period_suffix}"
             cos_feature = f"datetime_{column}_cos{period_suffix}"
-            df[sin_feature] = np.sin(2 * np.pi * df[column] / period)
-            df[cos_feature] = np.cos(2 * np.pi * df[column] / period)
-            self.generated_features.append(sin_feature)
-            self.generated_features.append(cos_feature)
+            if sin_feature not in df.columns:
+                df[sin_feature] = np.sin(2 * np.pi * df[column] / period)
+                self.generated_features.append(sin_feature)
+            if cos_feature not in df.columns:
+                df[cos_feature] = np.cos(2 * np.pi * df[column] / period)
+                self.generated_features.append(cos_feature)
 
         df["quarter"] = df[self.date_column].dt.quarter
 

{upgini-1.2.29a5 → upgini-1.2.29a7}/src/upgini/utils/email_utils.py

@@ -38,8 +38,9 @@ class EmailDomainGenerator:
     def generate(self, df: pd.DataFrame) -> pd.DataFrame:
         for email_col in self.email_columns:
             domain_feature = email_col + self.DOMAIN_SUFFIX
-            df[domain_feature] = df[email_col].apply(self._email_to_domain)
-            self.generated_features.append(domain_feature)
+            if domain_feature not in df.columns:
+                df[domain_feature] = df[email_col].apply(self._email_to_domain).astype("string")
+                self.generated_features.append(domain_feature)
         return df
 
     @staticmethod

{upgini-1.2.29a5 → upgini-1.2.29a7}/src/upgini/utils/features_validator.py

@@ -2,6 +2,7 @@ import logging
 from logging import Logger
 from typing import Dict, List, Optional, Tuple
 
+import numpy as np
 import pandas as pd
 from pandas.api.types import is_integer_dtype, is_object_dtype, is_string_dtype
 
@@ -83,10 +84,21 @@ class FeaturesValidator:
         return [
             i
             for i in df
-            if (is_object_dtype(df[i]) or is_string_dtype(df[i]) or is_integer_dtype(df[i]))
+            if (is_object_dtype(df[i]) or is_string_dtype(df[i]) or FeaturesValidator.__is_integer(df[i]))
             and (df[i].nunique(dropna=False) / row_count >= 0.85)
         ]
 
+    @staticmethod
+    def __is_integer(series: pd.Series) -> bool:
+        return (
+            is_integer_dtype(series)
+            or series.dropna()
+            .apply(
+                lambda f: (float.is_integer(f) and abs(f) < np.iinfo(np.int64).max) if isinstance(f, float) else False
+            )
+            .all()
+        )
+
     @staticmethod
     def find_constant_features(df: pd.DataFrame) -> List[str]:
         return [i for i in df if df[i].nunique() <= 1]

upgini-1.2.29a5/src/upgini/__about__.py

@@ -1 +0,0 @@
-__version__ = "1.2.29a5"