в блокнот research добавлена часть о выбранной лучшей модели

research/research.py

@@ -742,15 +742,21 @@ repr(optuna_study.best_params)
 # %%
 regressor_best_params = dict(optuna_study.best_params.items())
 
+
 # %% [markdown]
 # Составной пайплайн:
 
 # %%
-pipeline = build_pipeline(
+def build_pipeline_optimized_best():
+    return build_pipeline(
         regressor_best_params['n_estimators'],
         regressor_max_depth=regressor_best_params['max_depth'],
         regressor_max_features=regressor_best_params['max_features'],
     )
+
+
+# %%
+pipeline = build_pipeline_optimized_best()
 pipeline
 
 # %%
@@ -804,4 +810,49 @@ mlflow_log_model(
     ),
 )
 
+# %% [markdown]
+# ### И в продакшн
+
+# %% [markdown]
+# Лучшая выбранная модель — с автоматически подобранными гиперпараметрами.
+
+# %%
+pipeline = build_pipeline_optimized_best()
+pipeline
+
+# %%
+model_params = filter_params(
+    pipeline.get_params(),
+    include={
+        'preprocess': (False, PREPROCESS_AUGMENTING_TRANSFORMER_PARAMS_COMMON_INCLUDE.copy()),
+        'select_features': (False, FEATURE_SELECTOR_PARAMS_COMMON_INCLUDE.copy()),
+        'regress': (False, True),
+    },
+    exclude={
+        'preprocess': PREPROCESS_AUGMENTING_TRANSFORMER_PARAMS_COMMON_EXCLUDE.copy(),
+        'select_features': FEATURE_SELECTOR_PARAMS_COMMON_EXCLUDE,
+        'regress': RANDOM_FOREST_REGRESSOR_PARAMS_COMMON_EXCLUDE,
+    },
+)
+model_params
+
+# %%
+_ = pipeline.fit(df_orig_features, df_target.iloc[:, 0])
+
+# %%
+mlflow_log_model(
+    pipeline,
+    model_params=model_params,
+    metrics=None,
+    nested_run_name='Final model',
+    model_signature=mlflow_model_signature,
+    input_example=df_orig_features.head(MODEL_INOUT_EXAMPLE_SIZE),
+    #pip_requirements=str(MODEL_PIP_REQUIREMENTS_PATH),
+    comment_file_path=(
+        model_comment_path
+        if model_comment_path is not None
+        else (BASE_PATH / 'research' / model_comment_relpath)
+    ),
+)
+
 # %%