move extra to parameters

template/configs/train_config.yaml

@@ -20,16 +20,12 @@ steps:
     parameters:
       name: {{ product_name }}
 {%- if metric_compare_promotion %}
-  compute_performance_metrics_on_current_data:
-    parameters:
-      target_env: {{ target_environment }}
   promote_with_metric_compare:
 {%- else %}
   promote_latest_version:
 {%- endif %}
     parameters:
       mlflow_model_name: {{ product_name }}
-      target_env: {{ target_environment }}
   notify_on_success:
     parameters:
       notify_on_success: False
@@ -56,6 +52,9 @@ model_version:
 # pipeline level extra configurations
 extra:
   notify_on_failure: True
+# pipeline level parameters
+parameters:
+  target_env: {{ target_environment }}
 {%- if hyperparameters_tuning %}
   # This set contains all the model configurations that you want 
   # to evaluate during hyperparameter tuning stage.

template/pipelines/training.py

@@ -1,7 +1,7 @@
 # {% include 'template/license_header' %}
 
 
-from typing import List, Optional
+from typing import List, Optional, Any, Dict
 import random
 
 from steps import (
@@ -23,7 +23,7 @@
     promote_latest_version,
 {%- endif %}
 )
-from zenml import pipeline, get_pipeline_context
+from zenml import pipeline
 from zenml.logger import get_logger
 {%- if hyperparameters_tuning %}
 
@@ -38,6 +38,12 @@
 
 @pipeline(on_failure=notify_on_failure)
 def {{product_name}}_training(
+{%- if hyperparameters_tuning %}
+    model_search_space: Dict[str,Any],
+{%- else %}
+    model_configuration: Dict[str,Any],
+{%- endif %}
+    target_env: str,
     test_size: float = 0.2,
     drop_na: Optional[bool] = None,
     normalize: Optional[bool] = None,
@@ -54,6 +60,12 @@ def {{product_name}}_training(
     trains and evaluates a model.
 
     Args:
+{%- if hyperparameters_tuning %}
+        model_search_space: Search space for hyperparameter tuning
+{%- else %}
+        model_configuration: Configuration of the model to train
+{%- endif %}
+        target_env: The environment to promote the model to
         test_size: Size of holdout set for training 0.0..1.0
         drop_na: If `True` NA values will be removed from dataset
         normalize: If `True` dataset will be normalized with MinMaxScaler
@@ -62,12 +74,10 @@ def {{product_name}}_training(
         min_test_accuracy: Threshold to stop execution if test set accuracy is lower
         fail_on_accuracy_quality_gates: If `True` and `min_train_accuracy` or `min_test_accuracy`
             are not met - execution will be interrupted early
-
     """
     ### ADD YOUR OWN CODE HERE - THIS IS JUST AN EXAMPLE ###
     # Link all the steps together by calling them and passing the output
     # of one step as the input of the next step.
-    pipeline_extra = get_pipeline_context().extra
     ########## ETL stage ##########
     raw_data, target, _ = data_loader(random_state=random.randint(0,100))
     dataset_trn, dataset_tst = train_data_splitter(
@@ -86,7 +96,7 @@ def {{product_name}}_training(
     ########## Hyperparameter tuning stage ##########
     after = []
     search_steps_prefix = "hp_tuning_search_"
-    for config_name,model_search_configuration in pipeline_extra["model_search_space"].items():
+    for config_name,model_search_configuration in model_search_space.items():
             step_name = f"{search_steps_prefix}{config_name}"
             hp_tuning_single_search(
                 id=step_name,
@@ -100,7 +110,6 @@ def {{product_name}}_training(
             after.append(step_name)
     best_model = hp_tuning_select_best_model(step_names=after, after=after)
 {%- else %}
-    model_configuration = pipeline_extra["model_configuration"]
     best_model = get_model_from_config(
         model_package=model_configuration["model_package"], 
         model_class=model_configuration["model_class"],
@@ -130,16 +139,18 @@ def {{product_name}}_training(
 {%- if metric_compare_promotion %}
     latest_metric,current_metric = compute_performance_metrics_on_current_data(
         dataset_tst=dataset_tst,
+        target_env=target_env,
         after=["model_evaluator"]
     )
 
     promote_with_metric_compare(
         latest_metric=latest_metric,
         current_metric=current_metric,
+        target_env=target_env,
     )
     last_step = "promote_with_metric_compare"
 {%- else %}
-    promote_latest_version(after=["model_evaluator"])
+    promote_latest_version(target_env=target_env,after=["model_evaluator"])
     last_step = "promote_latest_version"
 {%- endif %}