upgini 1.2.38a3769.dev11__tar.gz → 1.2.39__tar.gz

{upgini-1.2.38a3769.dev11 → upgini-1.2.39}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: upgini
-Version: 1.2.38a3769.dev11
+Version: 1.2.39
 Summary: Intelligent data search & enrichment for Machine Learning
 Project-URL: Bug Reports, https://github.com/upgini/upgini/issues
 Project-URL: Homepage, https://upgini.com/
@@ -382,6 +382,7 @@ enricher = FeaturesEnricher(
 	date_format = "%Y-%d-%m"
 )
 ```
+
 ### 4. 🔍 Start your first feature search!
 The main abstraction you interact is `FeaturesEnricher`, a Scikit-learn compatible estimator.  You can easily add it into your existing ML pipelines. 
 Create instance of the `FeaturesEnricher` class and call:
@@ -412,7 +413,7 @@ enricher = FeaturesEnricher(
 enricher.fit(X, y)
 ```
 
-That's all). We've fitted `FeaturesEnricher`. 
+That's all! We've fit `FeaturesEnricher`. 
 ### 5. 📈 Evaluate feature importances (SHAP values) from the search result
 
 `FeaturesEnricher` class has two properties for feature importances, which will be filled after fit - `feature_names_` and `feature_importances_`:  
@@ -464,7 +465,7 @@ enricher = FeaturesEnricher(
 ) 
 ```
 
-## 💻 How it works?
+## 💻 How does it work?
 
 ### 🧹 Search dataset validation
 We validate and clean search initialization dataset under the hood:  
@@ -506,6 +507,17 @@ enricher = FeaturesEnricher(
 	cv=CVType.time_series
 )
 ```
+
+If you're working with multivariate time series, you should specify id columns of individual univariate series in `FeaturesEnricher`. For example, if you have a dataset predicting sales for different stores and products, you should specify store and product id columns as follows:
+```python
+enricher = FeaturesEnricher(
+    search_keys={
+        "sales_date": SearchKey.DATE,
+    },
+    id_columns=["store_id", "product_id"],
+    cv=CVType.time_series
+)
+```
 ⚠️ **Pre-process search dataset** in case of time series prediction:  
 sort rows in dataset according to observation order, in most cases - ascending order by date/datetime.
 

{upgini-1.2.38a3769.dev11 → upgini-1.2.39}/README.md

@@ -340,6 +340,7 @@ enricher = FeaturesEnricher(
 	date_format = "%Y-%d-%m"
 )
 ```
+
 ### 4. 🔍 Start your first feature search!
 The main abstraction you interact is `FeaturesEnricher`, a Scikit-learn compatible estimator.  You can easily add it into your existing ML pipelines. 
 Create instance of the `FeaturesEnricher` class and call:
@@ -370,7 +371,7 @@ enricher = FeaturesEnricher(
 enricher.fit(X, y)
 ```
 
-That's all). We've fitted `FeaturesEnricher`. 
+That's all! We've fit `FeaturesEnricher`. 
 ### 5. 📈 Evaluate feature importances (SHAP values) from the search result
 
 `FeaturesEnricher` class has two properties for feature importances, which will be filled after fit - `feature_names_` and `feature_importances_`:  
@@ -422,7 +423,7 @@ enricher = FeaturesEnricher(
 ) 
 ```
 
-## 💻 How it works?
+## 💻 How does it work?
 
 ### 🧹 Search dataset validation
 We validate and clean search initialization dataset under the hood:  
@@ -464,6 +465,17 @@ enricher = FeaturesEnricher(
 	cv=CVType.time_series
 )
 ```
+
+If you're working with multivariate time series, you should specify id columns of individual univariate series in `FeaturesEnricher`. For example, if you have a dataset predicting sales for different stores and products, you should specify store and product id columns as follows:
+```python
+enricher = FeaturesEnricher(
+    search_keys={
+        "sales_date": SearchKey.DATE,
+    },
+    id_columns=["store_id", "product_id"],
+    cv=CVType.time_series
+)
+```
 ⚠️ **Pre-process search dataset** in case of time series prediction:  
 sort rows in dataset according to observation order, in most cases - ascending order by date/datetime.
 

upgini-1.2.39/src/upgini/__about__.py

@@ -0,0 +1 @@
+__version__ = "1.2.39"

{upgini-1.2.38a3769.dev11 → upgini-1.2.39}/src/upgini/dataset.py

@@ -77,6 +77,7 @@ class Dataset:  # (pd.DataFrame):
         unnest_search_keys: Optional[Dict[str, str]] = None,
         model_task_type: Optional[ModelTaskType] = None,
         cv_type: Optional[CVType] = None,
+        date_column: Optional[str] = None,
         id_columns: Optional[List[str]] = None,
         random_state: Optional[int] = None,
         rest_client: Optional[_RestClient] = None,
@@ -122,6 +123,7 @@ class Dataset:  # (pd.DataFrame):
         self.columns_renaming: Dict[str, str] = {}
         self.imbalanced: bool = False
         self.id_columns = id_columns
+        self.date_column = date_column
         if logger is not None:
             self.logger = logger
         else:
@@ -232,6 +234,7 @@ class Dataset:  # (pd.DataFrame):
                 target_column=target_column,
                 task_type=self.task_type,
                 cv_type=self.cv_type,
+                date_column=self.date_column,
                 id_columns=self.id_columns,
                 random_state=self.random_state,
                 sample_size=self.FORCE_SAMPLE_SIZE,

{upgini-1.2.38a3769.dev11 → upgini-1.2.39}/src/upgini/features_enricher.py

@@ -932,9 +932,7 @@ class FeaturesEnricher(TransformerMixin):
                 cat_features, search_keys_for_metrics = self._get_client_cat_features(
                     estimator, validated_X, self.search_keys
                 )
-                search_keys_for_metrics.extend(
-                    [c for c in self.id_columns or [] if c not in search_keys_for_metrics]
-                )
+                search_keys_for_metrics.extend([c for c in self.id_columns or [] if c not in search_keys_for_metrics])
                 self.logger.info(f"Search keys for metrics: {search_keys_for_metrics}")
 
                 prepared_data = self._prepare_data_for_metrics(
@@ -2010,7 +2008,7 @@ class FeaturesEnricher(TransformerMixin):
             trace_id = trace_id or uuid.uuid4()
             return search_task.get_progress(trace_id)
 
-    def get_transactional_transform_api(self):
+    def get_transactional_transform_api(self, only_online_sources=False):
         if self.api_key is None:
             raise ValidationError(self.bundle.get("transactional_transform_unregistered"))
         if self._search_task is None:
@@ -2068,7 +2066,7 @@ class FeaturesEnricher(TransformerMixin):
         api_example = f"""curl 'https://search.upgini.com/online/api/http_inference_trigger?search_id={search_id}' \\
     -H 'Authorization: {self.api_key}' \\
     -H 'Content-Type: application/json' \\
-    -d '{{"search_keys": {keys}{features_section}}}'"""
+    -d '{{"search_keys": {keys}{features_section}, "only_online_sources": {str(only_online_sources).lower()}}}'"""
         return api_example
 
     def _get_copy_of_runtime_parameters(self) -> RuntimeParameters:
@@ -2112,13 +2110,15 @@ class FeaturesEnricher(TransformerMixin):
                 return None, {c: c for c in X.columns}, []
 
             features_meta = self._search_task.get_all_features_metadata_v2()
-            online_api_features = [fm.name for fm in features_meta if fm.from_online_api]
+            online_api_features = [fm.name for fm in features_meta if fm.from_online_api and fm.shap_value > 0]
             if len(online_api_features) > 0:
                 self.logger.warning(
                     f"There are important features for transform, that generated by online API: {online_api_features}"
                 )
-                # TODO
-                raise Exception("There are features selected that are paid. Contact support (sales@upgini.com)")
+                msg = self.bundle.get("online_api_features_transform").format(online_api_features)
+                self.logger.warning(msg)
+                print(msg)
+                print(self.get_transactional_transform_api(only_online_sources=True))
 
             if not metrics_calculation:
                 transform_usage = self.rest_client.get_current_transform_usage(trace_id)
@@ -2150,6 +2150,9 @@ class FeaturesEnricher(TransformerMixin):
                 validated_X = validated_X.drop(columns=columns_to_drop)
 
             search_keys = self.search_keys.copy()
+            if self.id_columns is not None and self.cv is not None and self.cv.is_time_series():
+                self.search_keys.update({col: SearchKey.CUSTOM_KEY for col in self.id_columns})
+
             search_keys = self.__prepare_search_keys(
                 validated_X, search_keys, is_demo_dataset, is_transform=True, silent_mode=silent_mode
             )
@@ -2307,7 +2310,8 @@ class FeaturesEnricher(TransformerMixin):
                 meaning_types=meaning_types,
                 search_keys=combined_search_keys,
                 unnest_search_keys=unnest_search_keys,
-                id_columns=self.__get_renamed_id_columns(),
+                id_columns=self.__get_renamed_id_columns(columns_renaming),
+                date_column=self._get_date_column(search_keys),
                 date_format=self.date_format,
                 rest_client=self.rest_client,
                 logger=self.logger,
@@ -2794,6 +2798,7 @@ class FeaturesEnricher(TransformerMixin):
             model_task_type=self.model_task_type,
             cv_type=self.cv,
             id_columns=self.__get_renamed_id_columns(),
+            date_column=self._get_date_column(self.fit_search_keys),
             date_format=self.date_format,
             random_state=self.random_state,
             rest_client=self.rest_client,
@@ -2953,8 +2958,9 @@ class FeaturesEnricher(TransformerMixin):
     def __should_add_date_column(self):
         return self.add_date_if_missing or (self.cv is not None and self.cv.is_time_series())
 
-    def __get_renamed_id_columns(self):
-        reverse_renaming = {v: k for k, v in self.fit_columns_renaming.items()}
+    def __get_renamed_id_columns(self, renaming: Optional[Dict[str, str]] = None):
+        renaming = renaming or self.fit_columns_renaming
+        reverse_renaming = {v: k for k, v in renaming.items()}
         return None if self.id_columns is None else [reverse_renaming.get(c) or c for c in self.id_columns]
 
     def __adjust_cv(self, df: pd.DataFrame):
@@ -3265,6 +3271,7 @@ class FeaturesEnricher(TransformerMixin):
                 f"Generate features: {self.generate_features}\n"
                 f"Round embeddings: {self.round_embeddings}\n"
                 f"Detect missing search keys: {self.detect_missing_search_keys}\n"
+                f"Exclude columns: {self.exclude_columns}\n"
                 f"Exclude features sources: {exclude_features_sources}\n"
                 f"Calculate metrics: {calculate_metrics}\n"
                 f"Scoring: {scoring}\n"
@@ -3272,6 +3279,15 @@ class FeaturesEnricher(TransformerMixin):
                 f"Remove target outliers: {remove_outliers_calc_metrics}\n"
                 f"Exclude columns: {self.exclude_columns}\n"
                 f"Search id: {self.search_id}\n"
+                f"Custom loss: {self.loss}\n"
+                f"Logs enabled: {self.logs_enabled}\n"
+                f"Raise validation error: {self.raise_validation_error}\n"
+                f"Baseline score column: {self.baseline_score_column}\n"
+                f"Client ip: {self.client_ip}\n"
+                f"Client visitorId: {self.client_visitorid}\n"
+                f"Add date if missing: {self.add_date_if_missing}\n"
+                f"Select features: {self.select_features}\n"
+                f"Disable force downsampling: {self.disable_force_downsampling}\n"
             )
 
             def sample(df):
@@ -3955,7 +3971,7 @@ class FeaturesEnricher(TransformerMixin):
             display_html_dataframe(self.metrics, self.metrics, msg)
 
     def __show_selected_features(self, search_keys: Dict[str, SearchKey]):
-        search_key_names = search_keys.keys()
+        search_key_names = [col for col, tpe in search_keys.items() if tpe != SearchKey.CUSTOM_KEY]
         if self.fit_columns_renaming:
             search_key_names = [self.fit_columns_renaming.get(col, col) for col in search_key_names]
         msg = self.bundle.get("features_info_header").format(len(self.feature_names_), search_key_names)

{upgini-1.2.38a3769.dev11 → upgini-1.2.39}/src/upgini/resource_bundle/strings.properties

@@ -216,6 +216,7 @@ imbalanced_target=\nTarget is imbalanced and will be undersampled. Frequency of
 loss_selection_info=Using loss `{}` for feature selection
 loss_calc_metrics_info=Using loss `{}` for metrics calculation with default estimator
 forced_balance_undersample=For quick data retrieval, your dataset has been sampled. To use data search without data sampling please contact support (sales@upgini.com)
+online_api_features_transform=Please note that some of the selected features {} are provided through a slow enrichment interface and are not available via transformation. However, they can be accessed via the API:
 
 # Validation table
 validation_column_name_header=Column name

upgini-1.2.38a3769.dev11/src/upgini/__about__.py

@@ -1 +0,0 @@
-__version__ = "1.2.38a3769.dev11"