[Frontend] Support bad_words in the /v1/completions endpoint (#46793)

Signed-off-by: sungbin1015 <sbin@solbox.com>
Co-authored-by: Claude <noreply@anthropic.com>

tests/entrypoints/openai/completion/test_completion.py

@@ -9,6 +9,8 @@ import regex as re
 from openai import BadRequestError
 
 from tests.utils import RemoteOpenAIServer
+from vllm.entrypoints.openai.completion.protocol import CompletionRequest
+from vllm.sampling_params import SamplingParams
 from vllm.tokenizers import get_tokenizer
 
 # any model with a chat template should work here
@@ -730,3 +732,38 @@ async def test_invalid_grammar(client: openai.AsyncOpenAI, model_name: str):
                 "structured_outputs": {"grammar": invalid_simplified_sql_grammar}
             },
         )
+
+
+# Unit tests for bad_words in CompletionRequest.to_sampling_params()
+def test_completion_request_bad_words_to_sampling_params():
+    """bad_words should be forwarded to SamplingParams (parity with chat)."""
+    request = CompletionRequest(
+        model="test-model",
+        prompt="Hello",
+        bad_words=["foo", "bar"],
+        max_tokens=10,
+    )
+
+    sampling_params = request.to_sampling_params(
+        max_tokens=10,
+        default_sampling_params={},
+    )
+
+    assert isinstance(sampling_params, SamplingParams)
+    assert sampling_params.bad_words == ["foo", "bar"]
+
+
+def test_completion_request_bad_words_default_empty():
+    """bad_words defaults to an empty list, matching the chat endpoint."""
+    request = CompletionRequest(
+        model="test-model",
+        prompt="Hello",
+        max_tokens=10,
+    )
+
+    assert request.bad_words == []
+    sampling_params = request.to_sampling_params(
+        max_tokens=10,
+        default_sampling_params={},
+    )
+    assert sampling_params.bad_words == []

vllm/entrypoints/openai/completion/protocol.py

@@ -95,6 +95,7 @@ class CompletionRequest(OpenAIBaseModel):
     )
     allowed_token_ids: list[int] | None = None
     prompt_logprobs: int | None = None
+    bad_words: list[str] = Field(default_factory=list)
     # --8<-- [end:completion-sampling-params]
 
     # --8<-- [start:completion-extra-params]
@@ -371,6 +372,7 @@ class CompletionRequest(OpenAIBaseModel):
             structured_outputs=self.structured_outputs,
             logit_bias=self.logit_bias,
             allowed_token_ids=self.allowed_token_ids,
+            bad_words=self.bad_words,
             extra_args=extra_args or None,
             skip_clone=True,  # Created fresh per request, safe to skip clone
             repetition_detection=self.repetition_detection,