[Bugfix][Structured Output] Support outlines engine with reasoning outputs for DeepSeek R1 (#14114)

docs/source/features/reasoning_outputs.md

@@ -10,7 +10,9 @@ Reasoning models return a additional `reasoning_content` field in their outputs,
 
 vLLM currently supports the following reasoning models:
 
-- [DeepSeek R1 series](https://huggingface.co/collections/deepseek-ai/deepseek-r1-678e1e131c0169c0bc89728d) (`deepseek_r1`, which looks for `<think> ... </think>`)
+| Model Series | Parser Name | Structured Output Support |
+|--------------|-------------|------------------|
+| [DeepSeek R1 series](https://huggingface.co/collections/deepseek-ai/deepseek-r1-678e1e131c0169c0bc89728d) | `deepseek_r1` | `guided_json`, `guided_regex` |
 
 ## Quickstart
 
@@ -78,11 +80,51 @@ Streaming chat completions are also supported for reasoning models. The `reasoni
 
 Please note that it is not compatible with the OpenAI Python client library. You can use the `requests` library to make streaming requests. You could checkout the [example](https://github.com/vllm-project/vllm/blob/main/examples/online_serving/openai_chat_completion_with_reasoning_streaming.py).
 
+## Structured output
+
+The reasoning content is also available in the structured output. The structured output engine like `xgrammar` will use the reasoning content to generate structured output.
+
+```python
+from openai import OpenAI
+from pydantic import BaseModel
+
+# Modify OpenAI's API key and API base to use vLLM's API server.
+openai_api_key = "EMPTY"
+openai_api_base = "http://localhost:8000/v1"
+
+client = OpenAI(
+    api_key=openai_api_key,
+    base_url=openai_api_base,
+)
+
+models = client.models.list()
+model = models.data[0].id
+
+
+class People(BaseModel):
+    name: str
+    age: int
+
+
+json_schema = People.model_json_schema()
+
+prompt = ("Generate a JSON with the name and age of one random person.")
+completion = client.chat.completions.create(
+    model=model,
+    messages=[{
+        "role": "user",
+        "content": prompt,
+    }],
+    extra_body={"guided_json": json_schema},
+)
+print("reasoning_content: ", completion.choices[0].message.reasoning_content)
+print("content: ", completion.choices[0].message.content)
+```
+
 ## Limitations
 
 - The reasoning content is only available for online serving's chat completion endpoint (`/v1/chat/completions`).
 - It is not compatible with [`tool_calling`](#tool_calling).
-- The reasoning content is not available for all models. Check the model's documentation to see if it supports reasoning.
 
 ## How to support a new reasoning model
 
@@ -166,9 +208,10 @@ class DeepSeekReasoner(Reasoner):
 
     def is_reasoning_end(self, input_ids: list[int]) -> bool:
         return self.end_token_id in input_ids
+    ...
 ```
 
-The structured output engine like xgrammar will use `end_token_id` to check if the reasoning content is present in the model output and skip the structured output if it is the case.
+The structured output engine like `xgrammar` will use `end_token_id` to check if the reasoning content is present in the model output and skip the structured output if it is the case.
 
 Finally, you can enable reasoning for the model by using the `--enable-reasoning` and `--reasoning-parser` flags.
 

examples/online_serving/openai_chat_completion_structured_outputs_with_reasoning.py

@@ -33,6 +33,42 @@ client = OpenAI(
 models = client.models.list()
 model = models.data[0].id
 
+# Guided decoding by Regex
+prompt = ("What is the capital of France?")
+
+completion = client.chat.completions.create(
+    model=model,
+    messages=[{
+        "role": "user",
+        "content": prompt,
+    }],
+    extra_body={
+        "guided_regex": "(Paris|London)",
+    },
+)
+print("reasoning_content: ", completion.choices[0].message.reasoning_content)
+print("content: ", completion.choices[0].message.content)
+
+
+class People(BaseModel):
+    name: str
+    age: int
+
+
+json_schema = People.model_json_schema()
+
+prompt = ("Generate a JSON with the name and age of one random person.")
+completion = client.chat.completions.create(
+    model=model,
+    messages=[{
+        "role": "user",
+        "content": prompt,
+    }],
+    extra_body={"guided_json": json_schema},
+)
+print("reasoning_content: ", completion.choices[0].message.reasoning_content)
+print("content: ", completion.choices[0].message.content)
+
 
 # Guided decoding by JSON using Pydantic schema
 class CarType(str, Enum):
@@ -51,7 +87,7 @@ class CarDescription(BaseModel):
 json_schema = CarDescription.model_json_schema()
 
 prompt = ("Generate a JSON with the brand, model and car_type of"
-          "the most iconic car from the 90's, think in 100 tokens")
+          "the most iconic car from the 90's")
 completion = client.chat.completions.create(
     model=model,
     messages=[{
@@ -60,5 +96,34 @@ completion = client.chat.completions.create(
     }],
     extra_body={"guided_json": json_schema},
 )
-print("content", completion.choices[0].message.content)
 print("reasoning_content: ", completion.choices[0].message.reasoning_content)
+print("content: ", completion.choices[0].message.content)
+
+# Guided decoding by Grammar
+simplified_sql_grammar = """
+    ?start: select_statement
+
+    ?select_statement: "SELECT " column_list " FROM " table_name
+
+    ?column_list: column_name ("," column_name)*
+
+    ?table_name: identifier
+
+    ?column_name: identifier
+
+    ?identifier: /[a-zA-Z_][a-zA-Z0-9_]*/
+"""
+
+# This may be very slow https://github.com/vllm-project/vllm/issues/12122
+prompt = ("Generate an SQL query to show the 'username' and 'email'"
+          "from the 'users' table.")
+completion = client.chat.completions.create(
+    model=model,
+    messages=[{
+        "role": "user",
+        "content": prompt,
+    }],
+    extra_body={"guided_grammar": simplified_sql_grammar},
+)
+print("reasoning_content: ", completion.choices[0].message.reasoning_content)
+print("content: ", completion.choices[0].message.content)

vllm/model_executor/guided_decoding/__init__.py

@@ -112,6 +112,7 @@ async def get_guided_decoding_logits_processor(
     reasoner = get_reasoner(tokenizer, reasoning_backend)
 
     guided_params = maybe_backend_fallback(guided_params)
+
     # CFG grammar not supported by LMFE, so we use outlines instead
     if guided_params.backend_name == 'outlines':
         # NOTE: lazy import outlines to avoid https://github.com/vllm-project/vllm/issues/4193

vllm/model_executor/guided_decoding/outlines_logits_processors.py

@@ -43,7 +43,7 @@ class BaseLogitsProcessor:
 
     def __init__(self, guide: Guide, reasoner: Optional[Reasoner]):
         self._guide: Guide = guide
-        self._reasoner = reasoner
+        self._reasoner: Optional[Reasoner] = reasoner
         # CFGState is used for the FSM state for CFGGuide
         self._fsm_state: DefaultDict[int, Union[int,
                                                 CFGState]] = defaultdict(int)
@@ -54,10 +54,14 @@ class BaseLogitsProcessor:
 
         # Skip the structured logits processing if reasoning is not finished.
         # reasoner is not None only when `--enable-reasoning` is set.
-        if self._reasoner is not None and \
+        if self._reasoner is not None:
-        not self._reasoner.is_reasoning_end(
+            if not self._reasoner.is_reasoning_end(input_ids):
-                input_ids):
+                return scores
-            return scores
+            else:
+                # Remove the reasoning tokens from the input_ids
+                # We need this because our implementation relies on the
+                # hash of the input_ids to store the FSM state.
+                input_ids = self._reasoner.extract_content(input_ids)
 
         seq_id = hash(tuple(input_ids))
 

vllm/model_executor/guided_decoding/reasoner/__init__.py

@@ -4,10 +4,13 @@ from __future__ import annotations
 
 from transformers import PreTrainedTokenizer
 
+from vllm.logger import init_logger
 from vllm.model_executor.guided_decoding.reasoner.deepseek_reasoner import (  # noqa: E501
     DeepSeekReasoner)
 from vllm.model_executor.guided_decoding.reasoner.reasoner import Reasoner
 
+logger = init_logger(__name__)
+
 
 def get_reasoner(tokenizer: PreTrainedTokenizer,
                  reasoning_backend: str | None) -> Reasoner | None:
@@ -17,7 +20,12 @@ def get_reasoner(tokenizer: PreTrainedTokenizer,
     elif reasoning_backend == "deepseek_r1":
         return DeepSeekReasoner.from_tokenizer(tokenizer)
     else:
-        raise ValueError(f"Unknown reasoning backend '{reasoning_backend}'")
+        # Raise a warning for unknown reasoning backend and return None
+        # We cannot raise an error here because some reasoning models
+        # may not have a corresponding Reasoner class.
+        logger.warning("Unknown reasoning backend %s for structured outputs ",
+                       reasoning_backend)
+        return None
 
 
 __all__ = ["Reasoner", "get_reasoner"]

vllm/model_executor/guided_decoding/reasoner/deepseek_reasoner.py

@@ -26,3 +26,13 @@ class DeepSeekReasoner(Reasoner):
 
     def is_reasoning_end(self, input_ids: list[int]) -> bool:
         return self.end_token_id in input_ids
+
+    def extract_content(self, input_ids: list[int]) -> list[int]:
+        """
+        Extract the content after the end tokens
+        """
+        if self.end_token_id not in input_ids or \
+            input_ids.index(self.end_token_id) + 1 == len(input_ids):
+            return []
+        else:
+            return input_ids[input_ids.index(self.end_token_id) + 1:]

vllm/model_executor/guided_decoding/reasoner/reasoner.py

@@ -17,3 +17,7 @@ class Reasoner(ABC):
     @abstractmethod
     def is_reasoning_end(self, input_ids: list[int]) -> bool:
         pass
+
+    @abstractmethod
+    def extract_content(self, input_ids: list[int]) -> list[int]:
+        pass

vllm/model_executor/guided_decoding/xgrammar_decoding.py

@@ -392,7 +392,7 @@ class XGrammarLogitsProcessor:
     def clone(self) -> XGrammarLogitsProcessor:
         """Create a new instance with shared compiled grammar
           but separate state"""
-        new_processor = XGrammarLogitsProcessor(self.config)
+        new_processor = XGrammarLogitsProcessor(self.config, self.reasoner)
 
         # Share the compiled grammar context (immutable after compilation)
         new_processor.ctx = self.ctx