Speed up dag parsing for google gen_ai_generative_model dag (#59587)

Author: amoghrajesh

providers/google/tests/system/google/cloud/gen_ai/example_gen_ai_generative_model.py

@@ -26,7 +26,6 @@
 from datetime import datetime
 
 import requests
-from vertexai.preview.evaluation import MetricPromptTemplateExamples
 
 try:
     from airflow.sdk import task
@@ -166,17 +165,30 @@ def _get_actual_models(key) -> dict[str, str]:
         "Baking a decadent chocolate cake requires creaming butter and sugar, beating in eggs and alternating dry ingredients with buttermilk before baking until done.",
     ],
 }
-METRICS = [
-    MetricPromptTemplateExamples.Pointwise.SUMMARIZATION_QUALITY,
-    MetricPromptTemplateExamples.Pointwise.GROUNDEDNESS,
-    MetricPromptTemplateExamples.Pointwise.VERBOSITY,
-    MetricPromptTemplateExamples.Pointwise.INSTRUCTION_FOLLOWING,
-    "exact_match",
-    "bleu",
-    "rouge_1",
-    "rouge_2",
-    "rouge_l_sum",
-]
+
+
+def _get_metrics():
+    """
+    Lazily import and return the metrics list.
+
+    This avoids slow imports during DAG parsing by deferring the import
+    until the operator is actually created.
+    """
+    from vertexai.preview.evaluation import MetricPromptTemplateExamples
+
+    return [
+        MetricPromptTemplateExamples.Pointwise.SUMMARIZATION_QUALITY,
+        MetricPromptTemplateExamples.Pointwise.GROUNDEDNESS,
+        MetricPromptTemplateExamples.Pointwise.VERBOSITY,
+        MetricPromptTemplateExamples.Pointwise.INSTRUCTION_FOLLOWING,
+        "exact_match",
+        "bleu",
+        "rouge_1",
+        "rouge_2",
+        "rouge_l_sum",
+    ]
+
+
 EXPERIMENT_NAME = f"eval-test-experiment-airflow-operator-{ENV_ID}".replace("_", "-")
 EXPERIMENT_RUN_NAME = f"eval-experiment-airflow-operator-run-{ENV_ID}".replace("_", "-")
 PROMPT_TEMPLATE = "{instruction}. Article: {context}. Summary:"
@@ -281,7 +293,7 @@ def get_actual_models(key):
         location=REGION,
         pretrained_model=MULTIMODAL_MODEL,
         eval_dataset=EVAL_DATASET,
-        metrics=METRICS,
+        metrics=_get_metrics(),
         experiment_name=EXPERIMENT_NAME,
         experiment_run_name=EXPERIMENT_RUN_NAME,
         prompt_template=PROMPT_TEMPLATE,