[MNT] add cross feature engineering example

examples/dataset_table_workflow/base.py

@@ -149,10 +149,10 @@ class TableWorkflow:
                 #     recorder.record(user, idx, scores)
                 #     bar.update(1)
 
-                process_single_aug(user, idx, all_scores, recorders, save_root_path)
+                # process_single_aug(user, idx, all_scores, recorders, save_root_path)
                 recorder.save(save_path)
-            else:
-                process_single_aug(user, idx, recorder.data[user][idx], recorders, save_root_path)  
+            #else:
+            process_single_aug(user, idx, recorder.data[user][idx], recorders, save_root_path)  
         else:
             if test_info["force"] or recorder.should_test_method(user, idx, save_path):
                 scores = self._limited_data(method, test_info, loss_func)

examples/dataset_table_workflow/config.py

@@ -29,8 +29,8 @@ styles = {
 
 labels = {
     'user_model': "User Model",  
-    'single_aug': "Single Learnware Reuse (Select)",
-    "select_score": "Single Learnware Reuse (Select)",
+    'single_aug': "Single Learnware Reuse (FeatAug)",
+    "select_score": "Single Learnware Reuse (FeatAug)",
     'multiple_aug': "Multiple Learnware Reuse (FeatAug)",
     'ensemble_pruning': "Multiple Learnware Reuse (EnsemblePrune)",
     'multiple_avg': "Multiple Learnware Reuse (Averaging)"
@@ -95,6 +95,25 @@ user_model_params = {
             "MAX_ROUNDS": 1000,
             "early_stopping_rounds": 1000
         }
+    },
+    "PFS": {
+        "lgb": {
+            "params": {
+                "boosting_type": "gbdt",
+                "num_leaves": 2**7 - 1,
+                "learning_rate": 0.01,
+                "objective": "rmse",
+                "metric": "rmse",
+                "feature_fraction": 0.75,
+                "bagging_fraction": 0.75,
+                "bagging_freq": 5,
+                "seed": 1,
+                "verbose": -100,
+                "n_estimators": 100000,
+            },
+            "MAX_ROUNDS": 1000,
+            "early_stopping_rounds": 1000
+        }
     }
 }
 
@@ -126,11 +145,77 @@ hetero_cross_feat_eng_benchmark_config = BenchmarkConfig(
     name="PFS",
     user_num=41,
     learnware_ids = [
-        
+        "00000394", "00000393", "00000392", "00000391",
+        "00000390", "00000389", "00000388", "00000387",
+        "00000386", "00000385", "00000384", "00000383",
+        "00000382", "00000381", "00000380", "00000379",
+        "00000378", "00000377", "00000376", "00000375",
+        "00000374", "00000373", "00000372", "00000371",
+        "00000370", "00000369", "00000368", "00000367",
+        "00000366", "00000365", "00000364", "00000363",
+        "00000362", "00000361", "00000360", "00000359",
+        "00000358", "00000357", "00000356", "00000355",
+        "00000354", "00000353", "00000352", "00000351",
+        "00000350", "00000349", "00000348", "00000347",
+        "00000346", "00000345", "00000344", "00000343",
+        "00000342", "00000444", "00000443", "00000442",
+        "00000441", "00000440", "00000439", "00000438",
+        "00000437", "00000436", "00000435", "00000434",
+        "00000433", "00000432", "00000431", "00000430",
+        "00000429", "00000428", "00000427", "00000426",
+        "00000425", "00000424", "00000423", "00000422",
+        "00000421", "00000420", "00000419", "00000418",
+        "00000417", "00000416", "00000415", "00000414",
+        "00000413", "00000412", "00000411", "00000410",
+        "00000409", "00000408", "00000407", "00000406",
+        "00000405", "00000404", "00000403", "00000402",
+        "00000401", "00000400", "00000399", "00000398",
+        "00000397", "00000396", "00000395", "00000783",
+        "00000782", "00000781", "00000780", "00000779",
+        "00000778", "00000777", "00000776", "00000775",
+        "00000774", "00000773", "00000772", "00000771",
+        "00000770", "00000769", "00000768", "00000767",
+        "00000766", "00000765", "00000764", "00000763",
+        "00000762", "00000761", "00000760", "00000759",
+        "00000758", "00000757", "00000756", "00000755",
+        "00000754", "00000753", "00000752", "00000751",
+        "00000750", "00000749", "00000748", "00000747",
+        "00000746", "00000745", "00000744", "00000743",
+        "00000742", "00000741", "00000740", "00000739",
+        "00000738", "00000737", "00000736", "00000735",
+        "00000734", "00000733", "00000732", "00000731",
+        "00000730", "00000839", "00000838", "00000837",
+        "00000836", "00000835", "00000834", "00000833",
+        "00000832", "00000831", "00000830", "00000829",
+        "00000828", "00000827", "00000826", "00000825",
+        "00000824", "00000823", "00000822", "00000821",
+        "00000820", "00000819", "00000818", "00000817",
+        "00000816", "00000815", "00000814", "00000813",
+        "00000812", "00000811", "00000810", "00000809",
+        "00000808", "00000807", "00000806", "00000805",
+        "00000804", "00000803", "00000802", "00000801",
+        "00000800", "00000799", "00000798", "00000797",
+        "00000796", "00000795", "00000794", "00000793",
+        "00000792", "00000791", "00000790", "00000789",
+        "00000788", "00000787", "00000786", "00000912",
+        "00000911", "00000910", "00000909", "00000908",
+        "00000907", "00000906", "00000905", "00000904",
+        "00000903", "00000902", "00000901", "00000900",
+        "00000899", "00000898", "00000897", "00000896",
+        "00000895", "00000894", "00000893", "00000892",
+        "00000891", "00000890", "00000889", "00000888",
+        "00000887", "00000886", "00000885", "00000884",
+        "00000883", "00000882", "00000881", "00000880",
+        "00000879", "00000878", "00000877", "00000876",
+        "00000875", "00000874", "00000873", "00000872",
+        "00000871", "00000870", "00000869", "00000868",
+        "00000867", "00000866", "00000865", "00000864",
+        "00000863", "00000862", "00000861", "00000860",
+        "00000859"
     ],
-    test_data_path=None,
-    train_data_path=None,
-    extra_info_path=None
+    test_data_path="PFS/test_data.zip",
+    train_data_path="PFS/train_data.zip",
+    extra_info_path="PFS/extra_info.zip"
 )
 
 hetero_cross_task_benchmark_config = BenchmarkConfig(

examples/dataset_table_workflow/hetero.py

@@ -158,7 +158,7 @@ class HeterogeneousDatasetWorkflow(TableWorkflow):
             logger.info(f"Hetero search result of user {user}_{idx}: mixture learnware num: {len(mixture_learnware_list)}")
 
             test_info = {"user": user, "idx": idx, "train_subsets": train_subsets, "test_x": test_x, "test_y": test_y, "n_labeled_list": hetero_n_labeled_list}
-            common_config = {"user_rkme": user_stat_spec, "multiple_learnwares": mixture_learnware_list}
+            common_config = {"user_rkme": user_stat_spec, "learnwares": mixture_learnware_list}
             method_configs = {
                 "user_model": {"dataset": self.benchmark.name, "model_type": "lgb"},
                 "hetero_single_aug": {"user_rkme": user_stat_spec, "learnwares": all_learnwares},

examples/dataset_table_workflow/workflow.py

@@ -29,7 +29,7 @@ class TableDatasetWorkflow:
         workflow = HeterogeneousDatasetWorkflow(
             benchmark_config=hetero_cross_feat_eng_benchmark_config,
             name="hetero",
-            rebuild=False
+            rebuild=True
         )
         workflow.unlabeled_hetero_table_example()