Update _map_finish_reason() and use it in span attribute as well

Author: aabmass

instrumentation-genai/opentelemetry-instrumentation-vertexai/src/opentelemetry/instrumentation/vertexai/utils.py

@@ -155,7 +155,8 @@ def get_genai_response_attributes(
     | prediction_service_v1beta1.GenerateContentResponse,
 ) -> dict[str, AttributeValue]:
     finish_reasons: list[str] = [
-        candidate.finish_reason.name for candidate in response.candidates
+        _map_finish_reason(candidate.finish_reason)
+        for candidate in response.candidates
     ]
     # TODO: add gen_ai.response.id once available in the python client
     # https://github.com/open-telemetry/opentelemetry-python-contrib/issues/3246
@@ -276,9 +277,5 @@ def _map_finish_reason(
     if finish_reason is EnumType.MAX_TOKENS:
         return "length"
 
-    # There are a lot of specific enum values from Vertex that would map to "content_filter".
-    # I'm worried trying to map the enum obfuscates the telemetry because 1) it over
-    # generalizes and 2) half of the values are from the OTel enum and others from the vertex
-    # enum. See for reference
-    # https://github.com/googleapis/python-aiplatform/blob/c5023698c7068e2f84523f91b824641c9ef2d694/google/cloud/aiplatform_v1/types/content.py#L786-L822
-    return finish_reason.name.lower()
+    # If there is no 1:1 mapping to an OTel preferred enum value, use the exact vertex reason
+    return finish_reason.name

instrumentation-genai/opentelemetry-instrumentation-vertexai/tests/test_chat_completions.py

@@ -38,7 +38,7 @@ def test_generate_content(
     assert dict(spans[0].attributes) == {
         "gen_ai.operation.name": "chat",
         "gen_ai.request.model": "gemini-1.5-flash-002",
-        "gen_ai.response.finish_reasons": ("STOP",),
+        "gen_ai.response.finish_reasons": ("stop",),
         "gen_ai.response.model": "gemini-1.5-flash-002",
         "gen_ai.system": "vertex_ai",
         "gen_ai.usage.input_tokens": 5,
@@ -106,7 +106,7 @@ def test_generate_content_without_events(
     assert dict(spans[0].attributes) == {
         "gen_ai.operation.name": "chat",
         "gen_ai.request.model": "gemini-1.5-flash-002",
-        "gen_ai.response.finish_reasons": ("STOP",),
+        "gen_ai.response.finish_reasons": ("stop",),
         "gen_ai.response.model": "gemini-1.5-flash-002",
         "gen_ai.system": "vertex_ai",
         "gen_ai.usage.input_tokens": 5,
@@ -294,7 +294,7 @@ def test_generate_content_extra_params(span_exporter, instrument_no_content):
         "gen_ai.request.stop_sequences": ("\n\n\n",),
         "gen_ai.request.temperature": 0.20000000298023224,
         "gen_ai.request.top_p": 0.949999988079071,
-        "gen_ai.response.finish_reasons": ("MAX_TOKENS",),
+        "gen_ai.response.finish_reasons": ("length",),
         "gen_ai.response.model": "gemini-1.5-flash-002",
         "gen_ai.system": "vertex_ai",
         "gen_ai.usage.input_tokens": 5,

instrumentation-genai/opentelemetry-instrumentation-vertexai/tests/test_utils.py

@@ -13,7 +13,19 @@
 # limitations under the License.
 
 
-from opentelemetry.instrumentation.vertexai.utils import get_server_attributes
+from opentelemetry.instrumentation.vertexai.utils import (
+    get_server_attributes,
+    _map_finish_reason,
+)
+import pytest
+from google.cloud.aiplatform_v1.types import (
+    content,
+    prediction_service,
+    tool,
+)
+from google.cloud.aiplatform_v1beta1.types import (
+    content as content_v1beta1,
+)
 
 
 def test_get_server_attributes() -> None:
@@ -30,3 +42,25 @@ def test_get_server_attributes() -> None:
         "server.address": "us-central1-aiplatform.googleapis.com",
         "server.port": 5432,
     }
+
+
+def test_map_finish_reason():
+    for Enum in (
+        content.Candidate.FinishReason,
+        content_v1beta1.Candidate.FinishReason,
+    ):
+        for finish_reason, expect in [
+            # Handled mappings
+            (Enum.FINISH_REASON_UNSPECIFIED, "error"),
+            (Enum.OTHER, "error"),
+            (Enum.STOP, "stop"),
+            (Enum.MAX_TOKENS, "length"),
+            # Preserve vertex enum value
+            (Enum.BLOCKLIST, "BLOCKLIST"),
+            (Enum.MALFORMED_FUNCTION_CALL, "MALFORMED_FUNCTION_CALL"),
+            (Enum.PROHIBITED_CONTENT, "PROHIBITED_CONTENT"),
+            (Enum.RECITATION, "RECITATION"),
+            (Enum.SAFETY, "SAFETY"),
+            (Enum.SPII, "SPII"),
+        ]:
+            assert _map_finish_reason(finish_reason) == expect