Fix non-ascii/utf8 encoding

ddtrace/llmobs/_llmobs.py

@@ -155,13 +155,13 @@ def _llmobs_span_event(cls, span: Span) -> Tuple[Dict[str, Any], bool]:
         if span_kind == "llm" and span._get_ctx_item(INPUT_MESSAGES) is not None:
             meta["input"]["messages"] = span._get_ctx_item(INPUT_MESSAGES)
         if span._get_ctx_item(INPUT_VALUE) is not None:
-            meta["input"]["value"] = safe_json(span._get_ctx_item(INPUT_VALUE))
+            meta["input"]["value"] = safe_json(span._get_ctx_item(INPUT_VALUE), ensure_ascii=False)
         if span_kind == "llm" and span._get_ctx_item(OUTPUT_MESSAGES) is not None:
             meta["output"]["messages"] = span._get_ctx_item(OUTPUT_MESSAGES)
         if span_kind == "embedding" and span._get_ctx_item(INPUT_DOCUMENTS) is not None:
             meta["input"]["documents"] = span._get_ctx_item(INPUT_DOCUMENTS)
         if span._get_ctx_item(OUTPUT_VALUE) is not None:
-            meta["output"]["value"] = safe_json(span._get_ctx_item(OUTPUT_VALUE))
+            meta["output"]["value"] = safe_json(span._get_ctx_item(OUTPUT_VALUE), ensure_ascii=False)
         if span_kind == "retrieval" and span._get_ctx_item(OUTPUT_DOCUMENTS) is not None:
             meta["output"]["documents"] = span._get_ctx_item(OUTPUT_DOCUMENTS)
         if span._get_ctx_item(INPUT_PROMPT) is not None:

ddtrace/llmobs/_utils.py

@@ -184,10 +184,10 @@ def _unserializable_default_repr(obj):
     return default_repr
 
 
-def safe_json(obj):
+def safe_json(obj, ensure_ascii=True):
     if isinstance(obj, str):
         return obj
     try:
-        return json.dumps(obj, ensure_ascii=False, skipkeys=True, default=_unserializable_default_repr)
+        return json.dumps(obj, ensure_ascii=ensure_ascii, skipkeys=True, default=_unserializable_default_repr)
     except Exception:
         log.error("Failed to serialize object to JSON.", exc_info=True)

ddtrace/llmobs/_writer.py

@@ -250,7 +250,8 @@ def __init__(
         headers = {}
         clients = []  # type: List[WriterClientBase]
         if is_agentless:
-            assert agentless_url, "agentless_url is required for agentless mode"
+            if not agentless_url:
+                raise ValueError("agentless_url is required for agentless mode")
             clients.append(LLMObsAgentlessEventClient())
             intake_url = agentless_url
             headers["DD-API-KEY"] = config._dd_api_key

releasenotes/notes/fix-llmobs-json-encoding-ascii-f8da77867a910de1.yaml

@@ -0,0 +1,4 @@
+---
+fixes:
+  - |
+    LLM Observability: This fix resolves an issue where annotating a span with non utf-8 input/output values resulted in encoding errors.

tests/llmobs/test_llmobs.py

@@ -247,18 +247,28 @@ def test_only_generate_span_events_from_llmobs_spans(tracer, llmobs_events):
     assert llmobs_events[0] == expected_grandchild_llmobs_span
 
 
-def test_utf_inputs_outputs(llmobs, llmobs_backend):
+def test_utf_non_ascii_io(llmobs, llmobs_backend):
+    with llmobs.workflow() as workflow_span:
+        with llmobs.llm(model_name="gpt-3.5-turbo-0125") as llm_span:
+            llmobs.annotate(llm_span, input_data="안녕, 지금 몇 시야?")
+            llmobs.annotate(workflow_span, input_data="안녕, 지금 몇 시야?")
+    events = llmobs_backend.wait_for_num_events(num=1)
+    assert len(events) == 1
+    assert events[0]["spans"][0]["meta"]["input"]["messages"][0]["content"] == "안녕, 지금 몇 시야?"
+    assert events[0]["spans"][1]["meta"]["input"]["value"] == "안녕, 지금 몇 시야?"
+
+
+def test_non_utf8_inputs_outputs(llmobs, llmobs_backend):
     """Test that latin1 encoded inputs and outputs are correctly decoded."""
     with llmobs.llm(model_name="gpt-3.5-turbo-0125") as span:
         llmobs.annotate(
             span,
-            # uncomment to repro issue
-            # input_data="The first Super Bowl, which was formally known as the First AFL–NFL World Championship Game, was played on January 15, 1967.",
-            input_data="The first Super Bowl, which was formally known as the First AFL-NFL World Championship Game, was played on January 15, 1967.",
+            input_data="The first Super Bowl (aka First AFL–NFL World Championship Game), was played in 1967.",
         )
 
     events = llmobs_backend.wait_for_num_events(num=1)
+    assert len(events) == 1
     assert (
         events[0]["spans"][0]["meta"]["input"]["messages"][0]["content"]
-        == "The first Super Bowl, which was formally known as the First AFL-NFL World Championship Game, was played on January 15, 1967."
+        == "The first Super Bowl (aka First AFL–NFL World Championship Game), was played in 1967."
     )