fix(openai): Avoid consuming iterables passed to the Completions API (#5489)

Avoid consuming single-use iterators passed to the Completions API.

All iterables that are not dictionaries or strings are transformed to lists in the internals of `openai` before they are passed to an API call.

Author: alexander-alderman-webb

sentry_sdk/integrations/openai.py

@@ -357,13 +357,6 @@ def _set_completions_api_input_data(
         _commmon_set_input_data(span, kwargs)
         return
 
-    system_instructions = _get_system_instructions_completions(messages)
-    if len(system_instructions) > 0:
-        span.set_data(
-            SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS,
-            json.dumps(_transform_system_instructions(system_instructions)),
-        )
-
     if isinstance(messages, str):
         normalized_messages = normalize_message_roles([messages])  # type: ignore
         scope = sentry_sdk.get_current_scope()
@@ -376,13 +369,29 @@ def _set_completions_api_input_data(
         _commmon_set_input_data(span, kwargs)
         return
 
+    # dict special case following https://github.com/openai/openai-python/blob/3e0c05b84a2056870abf3bd6a5e7849020209cc3/src/openai/_utils/_transform.py#L194-L197
+    if not isinstance(messages, Iterable) or isinstance(messages, dict):
+        set_data_normalized(span, SPANDATA.GEN_AI_OPERATION_NAME, "chat")
+        _commmon_set_input_data(span, kwargs)
+        return
+
+    messages = list(messages)
+    kwargs["messages"] = messages
+
+    system_instructions = _get_system_instructions_completions(messages)
+    if len(system_instructions) > 0:
+        span.set_data(
+            SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS,
+            json.dumps(_transform_system_instructions(system_instructions)),
+        )
+
     non_system_messages = [
         message
         for message in messages
         if not _is_system_instruction_completions(message)
     ]
     if len(non_system_messages) > 0:
-        normalized_messages = normalize_message_roles(non_system_messages)  # type: ignore
+        normalized_messages = normalize_message_roles(non_system_messages)
         scope = sentry_sdk.get_current_scope()
         messages_data = truncate_and_annotate_messages(normalized_messages, span, scope)
         if messages_data is not None:

tests/integrations/openai/test_openai.py

@@ -204,6 +204,21 @@ def test_nonstreaming_chat_completion_no_prompts(
             ],
             id="parts",
         ),
+        pytest.param(
+            iter(
+                [
+                    {
+                        "role": "system",
+                        "content": [
+                            {"type": "text", "text": "You are a helpful assistant."},
+                            {"type": "text", "text": "Be concise and clear."},
+                        ],
+                    },
+                    {"role": "user", "content": "hello"},
+                ]
+            ),
+            id="iterator",
+        ),
     ],
 )
 def test_nonstreaming_chat_completion(sentry_init, capture_events, messages, request):
@@ -335,6 +350,21 @@ async def test_nonstreaming_chat_completion_async_no_prompts(
             ],
             id="parts",
         ),
+        pytest.param(
+            iter(
+                [
+                    {
+                        "role": "system",
+                        "content": [
+                            {"type": "text", "text": "You are a helpful assistant."},
+                            {"type": "text", "text": "Be concise and clear."},
+                        ],
+                    },
+                    {"role": "user", "content": "hello"},
+                ]
+            ),
+            id="iterator",
+        ),
     ],
 )
 async def test_nonstreaming_chat_completion_async(
@@ -521,6 +551,21 @@ def test_streaming_chat_completion_no_prompts(
             ],
             id="parts",
         ),
+        pytest.param(
+            iter(
+                [
+                    {
+                        "role": "system",
+                        "content": [
+                            {"type": "text", "text": "You are a helpful assistant."},
+                            {"type": "text", "text": "Be concise and clear."},
+                        ],
+                    },
+                    {"role": "user", "content": "hello"},
+                ]
+            ),
+            id="iterator",
+        ),
     ],
 )
 def test_streaming_chat_completion(sentry_init, capture_events, messages, request):
@@ -757,6 +802,21 @@ async def test_streaming_chat_completion_async_no_prompts(
             ],
             id="parts",
         ),
+        pytest.param(
+            iter(
+                [
+                    {
+                        "role": "system",
+                        "content": [
+                            {"type": "text", "text": "You are a helpful assistant."},
+                            {"type": "text", "text": "Be concise and clear."},
+                        ],
+                    },
+                    {"role": "user", "content": "hello"},
+                ]
+            ),
+            id="iterator",
+        ),
     ],
 )
 async def test_streaming_chat_completion_async(