ISSUE-759: Add labels and annotations into DSL

kale/processors/nbprocessor.py

@@ -507,6 +507,13 @@ def parse_cell_metadata(self, metadata):
                 )
             parsed_tags["annotations"] = cell_annotations
 
+        if cell_labels:
+            if missing_step_names:
+                raise ValueError(
+                    "A cell can not provide Pod labels in a cell that does not declare a step name."
+                )
+            parsed_tags["labels"] = cell_labels
+
         if cell_limits:
             if missing_step_names:
                 raise ValueError(

kale/templates/pipeline_template.jinja2

@@ -1,7 +1,7 @@
 import json
 import kfp.dsl as kfp_dsl
 from kfp.dsl import Input, Output, Dataset, HTML, Metrics, Artifact, Model
-from kfp.kubernetes import security_context
+from kfp.kubernetes import add_pod_annotation, add_pod_label, security_context
 
 {{ lightweight_components | join('\n\n') }}
 
@@ -65,6 +65,18 @@ def auto_generated_pipeline(
     {%- endfor %}
     {%- endif %}
 
+    {%- if step.config.labels %}
+    {%- for label_key, label_value in step.config.labels.items() %}
+    add_pod_label(task={{ step.name }}_task, label_key="{{ label_key }}", label_value="{{ label_value }}")
+    {%- endfor %}
+    {%- endif %}
+
+    {%- if step.config.annotations %}
+    {%- for ann_key, ann_value in step.config.annotations.items() %}
+    add_pod_annotation(task={{ step.name }}_task, annotation_key="{{ ann_key }}", annotation_value="{{ ann_value }}")
+    {%- endfor %}
+    {%- endif %}
+
 {% endfor %}
 
 if __name__ == "__main__":

kale/tests/assets/kfp_dsl/iris.py

@@ -1,7 +1,7 @@
 import json
 import kfp.dsl as kfp_dsl
 from kfp.dsl import Input, Output, Dataset, HTML, Metrics, Artifact, Model
-from kfp.kubernetes import security_context
+from kfp.kubernetes import add_pod_annotation, add_pod_label, security_context
 
 
 @kfp_dsl.component(
@@ -304,6 +304,8 @@ def auto_generated_pipeline(
 
     load_transform_data_task.set_display_name("load-transform-data-step")
     load_transform_data_task.set_caching_options(enable_caching=False)
+    add_pod_label(task=load_transform_data_task,
+                  label_key="access-ml-pipeline", label_value="true")
 
     train_model_task = train_model_step(
         x_trn_input_artifact=load_transform_data_task.outputs["x_trn_output_artifact"],
@@ -325,6 +327,8 @@ def auto_generated_pipeline(
 
     train_model_task.set_display_name("train-model-step")
     train_model_task.set_caching_options(enable_caching=False)
+    add_pod_label(task=train_model_task,
+                  label_key="access-ml-pipeline", label_value="true")
 
     evaluate_model_task = evaluate_model_step(
         model_input_artifact=train_model_task.outputs["model_output_artifact"],
@@ -348,6 +352,8 @@ def auto_generated_pipeline(
 
     evaluate_model_task.set_display_name("evaluate-model-step")
     evaluate_model_task.set_caching_options(enable_caching=False)
+    add_pod_label(task=evaluate_model_task,
+                  label_key="access-ml-pipeline", label_value="true")
 
 
 if __name__ == "__main__":

kale/tests/assets/kfp_dsl/pipeline_parameters_and_metrics.py

@@ -1,7 +1,7 @@
 import json
 import kfp.dsl as kfp_dsl
 from kfp.dsl import Input, Output, Dataset, HTML, Metrics, Artifact, Model
-from kfp.kubernetes import security_context
+from kfp.kubernetes import add_pod_annotation, add_pod_label, security_context
 
 
 @kfp_dsl.component(