[CI] Fix mypy for `vllm/v1/structured_output` (#32722)

Signed-off-by: yewentao256 <zhyanwentao@126.com>

tools/pre_commit/mypy.py

@@ -49,6 +49,7 @@ FILES = [
     "vllm/v1/metrics",
     "vllm/v1/pool",
     "vllm/v1/sample",
+    "vllm/v1/structured_output",
     "vllm/v1/worker",
 ]
 
@@ -64,7 +65,6 @@ SEPARATE_GROUPS = [
     # v1 related
     "vllm/v1/kv_offload",
     "vllm/v1/spec_decode",
-    "vllm/v1/structured_output",
 ]
 
 # TODO(woosuk): Include the code from Megatron and HuggingFace.

vllm/reasoning/abs_reasoning_parsers.py

@@ -51,7 +51,7 @@ class ReasoningParser:
         return self.model_tokenizer.get_vocab()
 
     @abstractmethod
-    def is_reasoning_end(self, input_ids: list[int]) -> bool:
+    def is_reasoning_end(self, input_ids: Sequence[int]) -> bool:
         """
         Check if the reasoning content ends in the input_ids.
 
@@ -68,7 +68,7 @@ class ReasoningParser:
         """
 
     def is_reasoning_end_streaming(
-        self, input_ids: list[int], delta_ids: list[int]
+        self, input_ids: Sequence[int], delta_ids: Sequence[int]
     ) -> bool:
         """
         Check if the reasoning content ends in the input_ids on a

vllm/reasoning/basic_parsers.py

@@ -65,7 +65,7 @@ class BaseThinkingReasoningParser(ReasoningParser):
                 "think start/end tokens in the tokenizer!"
             )
 
-    def is_reasoning_end(self, input_ids: list[int]) -> bool:
+    def is_reasoning_end(self, input_ids: Sequence[int]) -> bool:
         start_token_id = self.start_token_id
         end_token_id = self.end_token_id
 
@@ -77,7 +77,7 @@ class BaseThinkingReasoningParser(ReasoningParser):
         return False
 
     def is_reasoning_end_streaming(
-        self, input_ids: list[int], delta_ids: list[int]
+        self, input_ids: Sequence[int], delta_ids: Sequence[int]
     ) -> bool:
         end_token_id = self.end_token_id
         return end_token_id in delta_ids

vllm/reasoning/deepseek_v3_reasoning_parser.py

@@ -41,7 +41,7 @@ class DeepSeekV3ReasoningParser(ReasoningParser):
         return self._parser.is_reasoning_end(input_ids)
 
     def is_reasoning_end_streaming(
-        self, input_ids: list[int], delta_ids: list[int]
+        self, input_ids: Sequence[int], delta_ids: Sequence[int]
     ) -> bool:
         return self._parser.is_reasoning_end_streaming(input_ids, delta_ids)
 

vllm/reasoning/gptoss_reasoning_parser.py

@@ -78,7 +78,7 @@ class GptOssReasoningParser(ReasoningParser):
         self.reasoning_end_token_ids_suffix = self.model_tokenizer.encode("<|message|>")
         self.reasoning_max_num_between_tokens = 20
 
-    def is_reasoning_end(self, input_ids: list[int]) -> bool:
+    def is_reasoning_end(self, input_ids: Sequence[int]) -> bool:
         end_token_ids_prefix = self.reasoning_end_token_ids_prefix
         end_token_ids_suffix = self.reasoning_end_token_ids_suffix
         assert len(end_token_ids_prefix) > 0, "reasoning_end_token_ids_prefix is empty"

vllm/reasoning/holo2_reasoning_parser.py

@@ -61,7 +61,7 @@ class Holo2ReasoningParser(ReasoningParser):
         return self._parser.is_reasoning_end(input_ids)
 
     def is_reasoning_end_streaming(
-        self, input_ids: list[int], delta_ids: list[int]
+        self, input_ids: Sequence[int], delta_ids: Sequence[int]
     ) -> bool:
         return self._parser.is_reasoning_end_streaming(input_ids, delta_ids)
 

vllm/reasoning/hunyuan_a13b_reasoning_parser.py

@@ -79,7 +79,7 @@ class HunyuanA13BReasoningParser(ReasoningParser):
         self.token_buffer = []
         self.text_buffer = ""
 
-    def is_reasoning_end(self, input_ids: list[int]) -> bool:
+    def is_reasoning_end(self, input_ids: Sequence[int]) -> bool:
         return self.current_state == "response"
 
     def extract_content_ids(self, input_ids: list[int]) -> list[int]:

vllm/reasoning/identity_reasoning_parser.py

@@ -31,12 +31,12 @@ class IdentityReasoningParser(ReasoningParser):
                 "constructor during construction."
             )
 
-    def is_reasoning_end(self, input_ids: list[int]) -> bool:
+    def is_reasoning_end(self, input_ids: Sequence[int]) -> bool:
         # Always return True, since we never treat reasoning specially
         return True
 
     def is_reasoning_end_streaming(
-        self, input_ids: list[int], delta_ids: list[int]
+        self, input_ids: Sequence[int], delta_ids: Sequence[int]
     ) -> bool:
         return True
 

vllm/reasoning/minimax_m2_reasoning_parser.py

@@ -88,7 +88,7 @@ class MiniMaxM2AppendThinkReasoningParser(ReasoningParser):
         super().__init__(tokenizer, *args, **kwargs)
         self.end_token_id = self.vocab.get("</think>")
 
-    def is_reasoning_end(self, input_ids: list[int]) -> bool:
+    def is_reasoning_end(self, input_ids: Sequence[int]) -> bool:
         end_token_id = self.end_token_id
         return any(input_id == end_token_id for input_id in reversed(input_ids))
 

vllm/reasoning/mistral_reasoning_parser.py

 # SPDX-License-Identifier: Apache-2.0
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
+from collections.abc import Sequence
 from functools import cached_property
 
 from vllm.entrypoints.openai.chat_completion.protocol import (
@@ -65,7 +66,7 @@ class MistralReasoningParser(BaseThinkingReasoningParser):
 
         return SpecialTokens.end_think
 
-    def is_reasoning_end(self, input_ids: list[int]) -> bool:
+    def is_reasoning_end(self, input_ids: Sequence[int]) -> bool:
         has_eot_token = False
 
         for id in input_ids[::-1]:

vllm/reasoning/olmo3_reasoning_parser.py

@@ -242,7 +242,7 @@ class Olmo3ReasoningParser(ReasoningParser):
             think_start=self.think_start, think_end=self.think_end
         )
 
-    def is_reasoning_end(self, input_ids: list[int]) -> bool:
+    def is_reasoning_end(self, input_ids: Sequence[int]) -> bool:
         text = self.model_tokenizer.decode(input_ids)
         return self.think_end in text
 

vllm/reasoning/step3_reasoning_parser.py

@@ -100,11 +100,11 @@ class Step3ReasoningParser(ReasoningParser):
 
             return reasoning, content
 
-    def is_reasoning_end(self, input_ids: list[int]) -> bool:
+    def is_reasoning_end(self, input_ids: Sequence[int]) -> bool:
         return self.think_end_token_id in input_ids
 
     def is_reasoning_end_streaming(
-        self, input_ids: list[int], delta_ids: list[int]
+        self, input_ids: Sequence[int], delta_ids: Sequence[int]
     ) -> bool:
         end_token_id = self.think_end_token_id
         return end_token_id in delta_ids

vllm/v1/attention/backends/fa_utils.py

 # SPDX-License-Identifier: Apache-2.0
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
+from typing import Any
+
 from vllm.logger import init_logger
 from vllm.platforms import current_platform
 
@@ -17,15 +19,15 @@ elif current_platform.is_xpu():
     from vllm._ipex_ops import ipex_ops
 
     reshape_and_cache_flash = ipex_ops.reshape_and_cache_flash
-    flash_attn_varlen_func = ipex_ops.flash_attn_varlen_func
-    get_scheduler_metadata = ipex_ops.get_scheduler_metadata
+    flash_attn_varlen_func = ipex_ops.flash_attn_varlen_func  # type: ignore[assignment]
+    get_scheduler_metadata = ipex_ops.get_scheduler_metadata  # type: ignore[assignment]
 
 elif current_platform.is_rocm():
     try:
-        from flash_attn import flash_attn_varlen_func  # noqa: F401
+        from flash_attn import flash_attn_varlen_func  # type: ignore[no-redef]
     except ImportError:
 
-        def flash_attn_varlen_func(*args, **kwargs):
+        def flash_attn_varlen_func(*args: Any, **kwargs: Any) -> Any:  # type: ignore[no-redef,misc]
             raise ImportError(
                 "ROCm platform requires upstream flash-attn "
                 "to be installed. Please install flash-attn first."

vllm/v1/attention/backends/mla/aiter_triton_mla.py

@@ -49,7 +49,7 @@ class AiterTritonMLAImpl(AiterMLAImpl):
     def _flash_attn_varlen_diff_headdims(
         self, q, k, v, return_softmax_lse=False, softmax_scale=None, **kwargs
     ):
-        result = self.flash_attn_varlen_func(
+        result = self.flash_attn_varlen_func(  # type: ignore[call-arg]
             q,
             k,
             v,

vllm/v1/attention/backends/mla/rocm_aiter_mla.py

@@ -230,7 +230,7 @@ class AiterMLAImpl(MLACommonImpl[AiterMLAMetadata]):
     def _flash_attn_varlen_diff_headdims(
         self, q, k, v, return_softmax_lse=False, softmax_scale=None, **kwargs
     ):
-        output = self.flash_attn_varlen_func(
+        output = self.flash_attn_varlen_func(  # type: ignore[call-arg]
             q=q,
             k=k,
             v=v,

vllm/v1/structured_output/__init__.py

@@ -294,7 +294,7 @@ class StructuredOutputManager:
             assert request.structured_output_request is not None
             if request.structured_output_request.reasoning_ended is None:
                 request.structured_output_request.reasoning_ended = (
-                    self.reasoner.is_reasoning_end(request.prompt_token_ids)
+                    self.reasoner.is_reasoning_end(request.prompt_token_ids or [])
                 )
             return request.structured_output_request.reasoning_ended
         return True
@@ -323,8 +323,9 @@ class StructuredOutputManager:
 
         # Check if reasoning ends in *this* step
         delta_from = request.num_computed_tokens - request.num_output_placeholders
+        all_token_ids = request.all_token_ids
         if self.reasoner.is_reasoning_end_streaming(
-            request.all_token_ids, request.all_token_ids[delta_from:]
+            all_token_ids, all_token_ids[delta_from:]
         ):
             # Reasoning just ended, so we shouldn't advance til
             # next pass

vllm/v1/structured_output/backend_guidance.py

@@ -284,6 +284,9 @@ def serialize_guidance_grammar(
 def validate_guidance_grammar(
     sampling_params: SamplingParams, tokenizer: llguidance.LLTokenizer | None = None
 ) -> None:
+    # if structured output is not enabled, there is nothing to validate
+    if sampling_params.structured_outputs is None:
+        return
     tp, grm = get_structured_output_key(sampling_params.structured_outputs)
     guidance_grm = serialize_guidance_grammar(tp, grm)
     err = llguidance.LLMatcher.validate_grammar(guidance_grm, tokenizer)

vllm/v1/structured_output/backend_xgrammar.py

@@ -69,7 +69,7 @@ class XgrammarBackend(StructuredOutputBackend):
                 if idx < vocab_size:
                     encoded_vocab[idx] = token
             stop_token_ids = [self.tokenizer.eos_token_id]
-            backend_str = self.tokenizer.tokenizer.backend_tokenizer.to_str()
+            backend_str = self.tokenizer.tokenizer.backend_tokenizer.to_str()  # type: ignore[attr-defined]
             metadata = xgr.TokenizerInfo._detect_metadata_from_hf(backend_str)
             tokenizer_info = xgr.TokenizerInfo(
                 encoded_vocab=encoded_vocab,