test(e2e): add basic string stop and matched stop test coverage for MLX

e2e_test/mlx/test_mlx_backend.py

@@ -13,6 +13,7 @@
 import platform
 import sys
 
+import openai
 import pytest
 
 pytestmark = pytest.mark.skipif(
@@ -21,6 +22,37 @@
 )
 
 
+def collect_streamed_completion(stream):
+    """Collect all text and the final choice from a streaming completion response."""
+    chunks = list(stream)
+    text = "".join(c.choices[0].text for c in chunks if c.choices and c.choices[0].text)
+    final_choice = next(
+        (c.choices[0] for c in reversed(chunks) if c.choices and c.choices[0].finish_reason), None
+    )
+    assert final_choice is not None, "No chunk with finish_reason found in stream"
+    return text, final_choice
+
+
+def assert_stop_text_trimmed(text, stop_text):
+    assert stop_text not in text, f"Stop text {stop_text!r} should not appear in output: {text!r}"
+
+
+def assert_matched_stop(choice, expected):
+    actual = getattr(choice, "matched_stop", None)
+    assert actual == expected, f"Expected matched_stop={expected!r}, got {actual!r}"
+
+
+def assert_api_error(err, expected_status: int, expected_code: str) -> None:
+    assert err.status_code == expected_status, (
+        f"Expected HTTP {expected_status}, got {err.status_code}"
+    )
+    body = getattr(err, "body", None) or {}
+    error_str = str(body) + str(getattr(err, "message", ""))
+    assert expected_code in error_str, (
+        f"Expected {expected_code!r} in error body, got: {error_str!r}"
+    )
+
+
 WEATHER_TOOL = {
     "type": "function",
     "function": {
@@ -60,6 +92,9 @@ class TestMlxBackend:
     # actual content within max_tokens.
     NO_THINKING = {"chat_template_kwargs": {"enable_thinking": False}}
 
+    STOP_SEQUENCE_TEST_PROMPT = "Repeat: 1 2 3 hello world 4 5 6 7"
+    SINGLE_STRING_STOP = "6"
+
     def test_basic_chat(self, model, api_client):
         response = api_client.chat.completions.create(
             model=model,
@@ -153,3 +188,66 @@ def test_max_tokens_finish_reason(self, model, api_client):
         )
         assert response.choices[0].finish_reason == "length"
         assert response.usage.completion_tokens == 10
+
+    def test_chat_stop_string_non_streaming(self, model, api_client):
+        response = api_client.chat.completions.create(
+            model=model,
+            messages=[{"role": "user", "content": self.STOP_SEQUENCE_TEST_PROMPT}],
+            max_tokens=160,
+            temperature=0,
+            stop=[self.SINGLE_STRING_STOP],
+            extra_body=self.NO_THINKING,
+        )
+        choice = response.choices[0]
+        assert choice.finish_reason == "stop"
+        assert_stop_text_trimmed(choice.message.content or "", self.SINGLE_STRING_STOP)
+        assert_matched_stop(choice, self.SINGLE_STRING_STOP)
+
+    def test_completion_stop_string_non_streaming(self, model, api_client):
+        response = api_client.completions.create(
+            model=model,
+            prompt=self.STOP_SEQUENCE_TEST_PROMPT,
+            max_tokens=160,
+            temperature=0,
+            stop=[self.SINGLE_STRING_STOP],
+        )
+        choice = response.choices[0]
+        assert choice.finish_reason == "stop"
+        assert_stop_text_trimmed(choice.text, self.SINGLE_STRING_STOP)
+        assert_matched_stop(choice, self.SINGLE_STRING_STOP)
+
+    def test_chat_rejects_multi_token_stop_string(self, model, api_client):
+        with pytest.raises((openai.BadRequestError, openai.APIStatusError)) as exc_info:
+            api_client.chat.completions.create(
+                model=model,
+                messages=[{"role": "user", "content": self.STOP_SEQUENCE_TEST_PROMPT}],
+                max_tokens=10,
+                stop=["hello world"],
+                extra_body=self.NO_THINKING,
+            )
+        assert_api_error(exc_info.value, 400, "unsupported_stop_string")
+
+    def test_completion_rejects_multi_token_stop_string(self, model, api_client):
+        with pytest.raises((openai.BadRequestError, openai.APIStatusError)) as exc_info:
+            api_client.completions.create(
+                model=model,
+                prompt=self.STOP_SEQUENCE_TEST_PROMPT,
+                max_tokens=10,
+                stop=["hello world"],
+            )
+        assert_api_error(exc_info.value, 400, "unsupported_stop_string")
+
+    def test_completion_stop_string_streaming_final_chunk(self, model, api_client):
+        """Streaming completion: final chunk has finish_reason='stop' and matched_stop set."""
+        stream = api_client.completions.create(
+            model=model,
+            prompt=self.STOP_SEQUENCE_TEST_PROMPT,
+            max_tokens=160,
+            temperature=0,
+            stop=[self.SINGLE_STRING_STOP],
+            stream=True,
+        )
+        text, final_choice = collect_streamed_completion(stream)
+        assert final_choice.finish_reason == "stop"
+        assert_stop_text_trimmed(text, self.SINGLE_STRING_STOP)
+        assert_matched_stop(final_choice, self.SINGLE_STRING_STOP)