Coerce top_k request attribute to integer across Bedrock, Portkey, and Smolagents (#707)

* Coerce top_k request attribute to integer across Bedrock, Portkey, and Smolagents

* Add changelog fragments

* Assert top_k integer type in Bedrock and Portkey tests

* Update InferenceInvocation.top_k type to int | None and add util test

Author: lmolkova

instrumentation/opentelemetry-instrumentation-genai-bedrock/.changelog/707.fixed

@@ -0,0 +1 @@
+Coerce top_k request parameter to integer.

instrumentation/opentelemetry-instrumentation-genai-bedrock/src/opentelemetry/instrumentation/genai/bedrock/extractors.py

@@ -307,7 +307,7 @@ def extract_converse_request(
         invocation.top_p = inf_config.get("topP")
         invocation.max_tokens = inf_config.get("maxTokens")
         invocation.stop_sequences = inf_config.get("stopSequences")
-        invocation.top_k = _safe_float(
+        invocation.top_k = _safe_int(
             _first_not_none(inf_config.get("topK"), inf_config.get("top_k"))
         )
         invocation.seed = inf_config.get("seed")
@@ -322,7 +322,7 @@ def extract_converse_request(
             add_inf.get("top_k") if _is_dict(add_inf) else None,
             invocation.top_k,
         )
-        invocation.top_k = _safe_float(top_k_val)
+        invocation.top_k = _safe_int(top_k_val)
         if "seed" in add_fields:
             invocation.seed = add_fields.get("seed")
 
@@ -496,7 +496,7 @@ def extract_invoke_model_request(
     )
 
     # Top K
-    invocation.top_k = _safe_float(
+    invocation.top_k = _safe_int(
         _first_not_none(
             body.get("top_k"),
             body.get("topK"),

instrumentation/opentelemetry-instrumentation-genai-bedrock/tests/test_converse.py

@@ -363,7 +363,8 @@ def test_extract_converse_request_top_k_and_seed(tracer_provider) -> None:
         },
         invocation,
     )
-    assert invocation.top_k == 40.0
+    assert invocation.top_k == 40
+    assert isinstance(invocation.top_k, int)
     assert invocation.seed == 123
 
     invocation2 = handler.inference(provider="aws.bedrock")
@@ -373,7 +374,8 @@ def test_extract_converse_request_top_k_and_seed(tracer_provider) -> None:
         },
         invocation2,
     )
-    assert invocation2.top_k == 250.0
+    assert invocation2.top_k == 250
+    assert isinstance(invocation2.top_k, int)
     assert invocation2.seed == 456
 
     invocation3 = handler.inference(provider="aws.bedrock")
@@ -383,7 +385,8 @@ def test_extract_converse_request_top_k_and_seed(tracer_provider) -> None:
         },
         invocation3,
     )
-    assert invocation3.top_k == 20.0
+    assert invocation3.top_k == 20
+    assert isinstance(invocation3.top_k, int)
 
     invocation4 = handler.inference(provider="aws.bedrock")
     extract_converse_request(
@@ -392,7 +395,8 @@ def test_extract_converse_request_top_k_and_seed(tracer_provider) -> None:
         },
         invocation4,
     )
-    assert invocation4.top_k == 0.0
+    assert invocation4.top_k == 0
+    assert isinstance(invocation4.top_k, int)
     assert invocation4.seed == 0
 
 

instrumentation/opentelemetry-instrumentation-genai-bedrock/tests/test_invoke_model.py

@@ -109,7 +109,10 @@ def test_invoke_model_anthropic_messages(
     assert span.attributes[GenAIAttributes.GEN_AI_REQUEST_MAX_TOKENS] == 100
     assert span.attributes[GenAIAttributes.GEN_AI_REQUEST_TEMPERATURE] == 0.7
     assert span.attributes[GenAIAttributes.GEN_AI_REQUEST_TOP_P] == 0.9
-    assert span.attributes[GenAIAttributes.GEN_AI_REQUEST_TOP_K] == 40.0
+    assert span.attributes[GenAIAttributes.GEN_AI_REQUEST_TOP_K] == 40
+    assert isinstance(
+        span.attributes[GenAIAttributes.GEN_AI_REQUEST_TOP_K], int
+    )
     assert span.attributes[GenAIAttributes.GEN_AI_RESPONSE_FINISH_REASONS] == (
         "stop",
     )
@@ -443,7 +446,8 @@ def test_extract_invoke_model_request_zero_values(tracer_provider) -> None:
     )
     assert invocation.temperature == 0.0
     assert invocation.top_p == 0.0
-    assert invocation.top_k == 0.0
+    assert invocation.top_k == 0
+    assert isinstance(invocation.top_k, int)
     assert invocation.max_tokens == 0
     assert invocation.seed == 0
 

instrumentation/opentelemetry-instrumentation-genai-portkey/.changelog/707.fixed

@@ -0,0 +1 @@
+Coerce top_k request parameter to integer.

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

@@ -250,7 +250,7 @@ def _apply_request_parameters(
     elif (p := get_value(kwargs.get("p"))) is not None:
         invocation.top_p = float(p)
     if (top_k := get_value(kwargs.get("top_k"))) is not None:
-        invocation.top_k = float(top_k)
+        invocation.top_k = int(top_k)
     if (max_tokens := get_value(kwargs.get("max_tokens"))) is not None:
         invocation.max_tokens = int(max_tokens)
     elif (

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

@@ -176,6 +176,7 @@ def test_sync_chat_completions_basic(
             model="gpt-4o",
             temperature=0.7,
             top_p=0.9,
+            top_k=40,
             max_tokens=150,
             seed=42,
         )
@@ -207,6 +208,10 @@ def test_sync_chat_completions_basic(
             == 0.7
         )
         assert span.attributes.get(GenAIAttributes.GEN_AI_REQUEST_TOP_P) == 0.9
+        assert span.attributes.get(GenAIAttributes.GEN_AI_REQUEST_TOP_K) == 40
+        assert isinstance(
+            span.attributes.get(GenAIAttributes.GEN_AI_REQUEST_TOP_K), int
+        )
         assert (
             span.attributes.get(GenAIAttributes.GEN_AI_REQUEST_MAX_TOKENS)
             == 150
@@ -598,13 +603,18 @@ def test_chat_completions_top_p_zero_and_max_tokens_zero(
             messages=[{"role": "user", "content": "Hi"}],
             model="gpt-4o",
             top_p=0.0,
+            top_k=0,
             max_tokens=0,
         )
 
         spans = span_exporter.get_finished_spans()
         assert len(spans) == 1
         span = spans[0]
         assert span.attributes.get(GenAIAttributes.GEN_AI_REQUEST_TOP_P) == 0.0
+        assert span.attributes.get(GenAIAttributes.GEN_AI_REQUEST_TOP_K) == 0
+        assert isinstance(
+            span.attributes.get(GenAIAttributes.GEN_AI_REQUEST_TOP_K), int
+        )
         assert (
             span.attributes.get(GenAIAttributes.GEN_AI_REQUEST_MAX_TOKENS) == 0
         )

instrumentation/opentelemetry-instrumentation-genai-smolagents/.changelog/707.fixed

@@ -0,0 +1 @@
+Coerce top_k request parameter to integer.

instrumentation/opentelemetry-instrumentation-genai-smolagents/src/opentelemetry/instrumentation/genai/smolagents/patch.py

@@ -224,7 +224,7 @@ def _apply_request_parameters(
 
     invocation.temperature = _coerce_float(merged.get("temperature"))
     invocation.top_p = _coerce_float(merged.get("top_p"))
-    invocation.top_k = _coerce_float(merged.get("top_k"))
+    invocation.top_k = _coerce_int(merged.get("top_k"))
     invocation.frequency_penalty = _coerce_float(
         merged.get("frequency_penalty")
     )

instrumentation/opentelemetry-instrumentation-genai-smolagents/tests/test_models.py

@@ -396,9 +396,8 @@ def test_request_parameters_recorded(
     (span,) = spans_by_operation(span_exporter.get_finished_spans(), "chat")
     assert attr(span, GenAI.GEN_AI_REQUEST_TEMPERATURE) == 0.5
     assert attr(span, GenAI.GEN_AI_REQUEST_TOP_P) == 0.9
-    # top_k is a float attribute in the spec even though callers pass an int.
-    assert attr(span, GenAI.GEN_AI_REQUEST_TOP_K) == 40.0
-    assert isinstance(attr(span, GenAI.GEN_AI_REQUEST_TOP_K), float)
+    assert attr(span, GenAI.GEN_AI_REQUEST_TOP_K) == 40
+    assert isinstance(attr(span, GenAI.GEN_AI_REQUEST_TOP_K), int)
     assert attr(span, GenAI.GEN_AI_REQUEST_FREQUENCY_PENALTY) == 0.25
     assert attr(span, GenAI.GEN_AI_REQUEST_PRESENCE_PENALTY) == 1.0
     assert attr(span, GenAI.GEN_AI_REQUEST_MAX_TOKENS) == 256