upgini 1.1.280.dev1__tar.gz → 1.1.282__tar.gz

{upgini-1.1.280.dev1 → upgini-1.1.282}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: upgini
-Version: 1.1.280.dev1
+Version: 1.1.282
 Summary: Intelligent data search & enrichment for Machine Learning
 Project-URL: Bug Reports, https://github.com/upgini/upgini/issues
 Project-URL: Homepage, https://upgini.com/
@@ -839,4 +839,4 @@ Some convenient ways to start contributing are:
 - [More perks for registered users](https://profile.upgini.com)
 
 <sup>😔 Found mistype or a bug in code snippet? Our bad! <a href="https://github.com/upgini/upgini/issues/new?assignees=&title=readme%2Fbug">
-Please report it here.</a></sup>
+Please report it here</a></sup>

{upgini-1.1.280.dev1 → upgini-1.1.282}/README.md

@@ -799,4 +799,4 @@ Some convenient ways to start contributing are:
 - [More perks for registered users](https://profile.upgini.com)
 
 <sup>😔 Found mistype or a bug in code snippet? Our bad! <a href="https://github.com/upgini/upgini/issues/new?assignees=&title=readme%2Fbug">
-Please report it here.</a></sup>
+Please report it here</a></sup>

upgini-1.1.282/src/upgini/__about__.py

@@ -0,0 +1 @@
+__version__ = "1.1.282"

{upgini-1.1.280.dev1 → upgini-1.1.282}/src/upgini/features_enricher.py

@@ -930,6 +930,7 @@ class FeaturesEnricher(TransformerMixin):
                         scoring,
                         groups=groups,
                         text_features=self.generate_features,
+                        has_date=has_date,
                     )
                     metric = wrapper.metric_name
                     multiplier = wrapper.multiplier
@@ -956,6 +957,7 @@ class FeaturesEnricher(TransformerMixin):
                             add_params=custom_loss_add_params,
                             groups=groups,
                             text_features=self.generate_features,
+                            has_date=has_date,
                         )
                         etalon_metric = baseline_estimator.cross_val_predict(
                             fitting_X, y_sorted, self.baseline_score_column
@@ -981,6 +983,7 @@ class FeaturesEnricher(TransformerMixin):
                             add_params=custom_loss_add_params,
                             groups=groups,
                             text_features=self.generate_features,
+                            has_date=has_date,
                         )
                         enriched_metric = enriched_estimator.cross_val_predict(fitting_enriched_X, enriched_y_sorted)
                         self.logger.info(f"Enriched {metric} on train combined features: {enriched_metric}")
@@ -1333,8 +1336,6 @@ class FeaturesEnricher(TransformerMixin):
         excluding_search_keys = list(search_keys.keys())
         if search_keys_for_metrics is not None and len(search_keys_for_metrics) > 0:
             excluding_search_keys = [sk for sk in excluding_search_keys if sk not in search_keys_for_metrics]
-        meta = self._search_task.get_all_features_metadata_v2()
-        zero_importance_client_features = [m.name for m in meta if m.source == "etalon" and m.shap_value == 0.0]
 
         client_features = [
             c
@@ -1344,7 +1345,6 @@ class FeaturesEnricher(TransformerMixin):
                 excluding_search_keys
                 + list(self.fit_dropped_features)
                 + [DateTimeSearchKeyConverter.DATETIME_COL, SYSTEM_RECORD_ID]
-                + zero_importance_client_features
             )
         ]
 
@@ -3720,7 +3720,7 @@ class FeaturesEnricher(TransformerMixin):
                     if y is not None:
                         with open(f"{tmp_dir}/y.pickle", "wb") as y_file:
                             pickle.dump(sample(y, xy_sample_index), y_file)
-                        if eval_set:
+                        if eval_set and _num_samples(eval_set[0][0]) > 0:
                             eval_xy_sample_index = rnd.randint(0, _num_samples(eval_set[0][0]), size=1000)
                             with open(f"{tmp_dir}/eval_x.pickle", "wb") as eval_x_file:
                                 pickle.dump(sample(eval_set[0][0], eval_xy_sample_index), eval_x_file)

{upgini-1.1.280.dev1 → upgini-1.1.282}/src/upgini/metrics.py

@@ -298,6 +298,7 @@ class EstimatorWrapper:
         scorer = check_scoring(self.estimator, scoring=self.scorer)
 
         if baseline_score_column is not None and self.metric_name == "GINI":
+            self.logger.info("Calculate baseline GINI on passed baseline_score_column and target")
             metric = roc_auc_score(y, x[baseline_score_column])
         else:
             cv_results = cross_validate(
@@ -314,9 +315,17 @@ class EstimatorWrapper:
             metrics_by_fold = cv_results["test_score"]
             self.cv_estimators = cv_results["estimator"]
 
+            self.check_fold_metrics(metrics_by_fold)
+
             metric = np.mean(metrics_by_fold) * self.multiplier
         return self.post_process_metric(metric)
 
+    def check_fold_metrics(self, metrics_by_fold: List[float]):
+        first_metric_sign = 1 if metrics_by_fold[0] >= 0 else -1
+        for metric in metrics_by_fold[1:]:
+            if first_metric_sign * metric < 0:
+                self.logger.warning(f"Sign of metrics differs between folds: {metrics_by_fold}")
+
     def post_process_metric(self, metric: float) -> float:
         if self.metric_name == "GINI":
             metric = 2 * metric - 1
@@ -346,6 +355,7 @@ class EstimatorWrapper:
         text_features: Optional[List[str]] = None,
         add_params: Optional[Dict[str, Any]] = None,
         groups: Optional[List[str]] = None,
+        has_date: Optional[bool] = None,
     ) -> EstimatorWrapper:
         scorer, metric_name, multiplier = _get_scorer(target_type, scoring)
         kwargs = {
@@ -360,6 +370,7 @@ class EstimatorWrapper:
         }
         if estimator is None:
             params = dict()
+            params["has_time"] = has_date
             # if metric_name.upper() in SUPPORTED_CATBOOST_METRICS:
             #     params["eval_metric"] = SUPPORTED_CATBOOST_METRICS[metric_name.upper()]
             if target_type == ModelTaskType.MULTICLASS:
@@ -475,7 +486,7 @@ class CatBoostWrapper(EstimatorWrapper):
 
         # Find rest categorical features
         self.cat_features = _get_cat_features(x, self.text_features, embedding_features)
-        x = fill_na_cat_features(x, self.cat_features)
+        # x = fill_na_cat_features(x, self.cat_features)
         unique_cat_features = []
         for name in self.cat_features:
             # Remove constant categorical features
@@ -525,7 +536,7 @@ class CatBoostWrapper(EstimatorWrapper):
             x, emb_columns = self.group_embeddings(x)
             params["embedding_features"] = emb_columns
         if self.cat_features:
-            x = fill_na_cat_features(x, self.cat_features)
+            # x = fill_na_cat_features(x, self.cat_features)
             params["cat_features"] = self.cat_features
 
         return x, y, params

{upgini-1.1.280.dev1 → upgini-1.1.282}/src/upgini/utils/display_utils.py

@@ -9,6 +9,7 @@ from typing import Callable, List, Optional
 
 import pandas as pd
 from xhtml2pdf import pisa
+from upgini.__about__ import __version__
 
 
 def ipython_available() -> bool:
@@ -166,12 +167,12 @@ def make_html_report(
                         /*-pdf-frame-border: 1;*/
                     }}
                     @frame content_frame {{
-                        left: 10pt; width: 574pt; top: 50pt; height: 752pt;
+                        left: 10pt; width: 574pt; top: 50pt; height: 742pt;
                         /*-pdf-frame-border: 1;*/
                     }}
                     @frame footer_frame {{
                         -pdf-frame-content: footer_content;
-                        left: 10pt; width: 574pt; top: 802pt; height: 30pt;
+                        left: 10pt; width: 574pt; top: 802pt; height: 40pt;
                         /*-pdf-frame-border: 1;*/
                     }}
                 }}
@@ -234,7 +235,8 @@ def make_html_report(
             <div id="header_content">UPGINI</div>
             <div id="footer_content">
                 © Upgini</br>
-                sales@upgini.com
+                sales@upgini.com</br>
+                Launched by version {__version__}
             </div>
 
             <h1>Data search report</h1>
@@ -257,7 +259,7 @@ def make_html_report(
             }
             <h3>Relevant data sources</h3>
             {make_table(relevant_datasources_df)}
-            <h3>All relevant features. Listing</h3>
+            <h3>All relevant features. Listing ({len(relevant_features_df)} items)</h3>
             {make_table(relevant_features_df, wrap_long_string=25)}
             {"<h3>Description of AutoFE feature names</h3>" + make_table(autofe_descriptions_df, wrap_long_string=25)
              if autofe_descriptions_df is not None

{upgini-1.1.280.dev1 → upgini-1.1.282}/src/upgini/utils/sklearn_ext.py

@@ -17,7 +17,7 @@ from sklearn.base import clone, is_classifier
 from sklearn.exceptions import FitFailedWarning, NotFittedError
 from sklearn.metrics import check_scoring
 from sklearn.metrics._scorer import _MultimetricScorer
-from sklearn.model_selection import check_cv
+from sklearn.model_selection import StratifiedKFold, check_cv
 from sklearn.utils.fixes import np_version, parse_version
 from sklearn.utils.validation import indexable
 
@@ -312,25 +312,34 @@ def cross_validate(
                 ret[key] = train_scores_dict[name]
 
         return ret
-    except Exception:
+    except ValueError as e:
         # logging.exception("Failed to execute overriden cross_validate. Fallback to original")
-        raise
-        # fit_params["use_best_model"] = False
-        # return original_cross_validate(
-        #     estimator,
-        #     X,
-        #     y,
-        #     groups=groups,
-        #     scoring=scoring,
-        #     cv=cv,
-        #     n_jobs=n_jobs,
-        #     verbose=verbose,
-        #     fit_params=fit_params,
-        #     pre_dispatch=pre_dispatch,
-        #     return_train_score=return_train_score,
-        #     return_estimator=return_estimator,
-        #     error_score=error_score,
-        # )
+        if hasattr(e, "args") and len(e.args) > 0 and "Only one class present in y_true" in e.args[0]:
+            # Try change CV to StratifiedKFold and retry
+            if hasattr(cv, "shuffle"):
+                shuffle = cv.shuffle
+            else:
+                shuffle = False
+            if hasattr(cv, "random_state"):
+                random_state = cv.random_state
+            else:
+                random_state = None
+            return cross_validate(
+                estimator,
+                x,
+                y,
+                groups=groups,
+                scoring=scoring,
+                cv=StratifiedKFold(n_splits=cv.get_n_splits(), shuffle=shuffle, random_state=random_state),
+                n_jobs=n_jobs,
+                verbose=verbose,
+                fit_params=fit_params,
+                pre_dispatch=pre_dispatch,
+                return_train_score=return_train_score,
+                return_estimator=return_estimator,
+                error_score=error_score,
+            )
+        raise e
 
 
 def _fit_and_score(

upgini-1.1.280.dev1/src/upgini/__about__.py

@@ -1 +0,0 @@
-__version__ = "1.1.280.dev1"