request.py 3.12 KB
Newer Older
1
# SPDX-License-Identifier: Apache-2.0
2
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
3
4
5
6
7
import dataclasses
import functools
import json
from concurrent.futures import Future
from concurrent.futures._base import TimeoutError
8
from typing import cast
9

10
from vllm.sampling_params import SamplingParams, StructuredOutputsParams
11
12
13
14
15
from vllm.v1.structured_output.backend_types import (
    StructuredOutputGrammar,
    StructuredOutputKey,
    StructuredOutputOptions,
)
16
17
18
19


@dataclasses.dataclass
class StructuredOutputRequest:
20
    params: StructuredOutputsParams
21
    _grammar: Future[StructuredOutputGrammar] | StructuredOutputGrammar | None = None
22
    reasoning_ended: bool | None = None
23

24
25
26
27
28
29
30
31
32
    @staticmethod
    def from_sampling_params(
        sampling_params: SamplingParams | None,
    ) -> "StructuredOutputRequest | None":
        if sampling_params is None:
            return None
        params = sampling_params.structured_outputs
        return StructuredOutputRequest(params=params) if params else None

33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
    def _check_grammar_completion(self) -> bool:
        # NOTE: We have to lazy import to gate circular imports
        from vllm.v1.request import RequestStatus

        if isinstance(self._grammar, Future):
            try:
                # We will check whether the future is ready within 100 us
                self._grammar = self._grammar.result(timeout=0.0001)
                self.status = RequestStatus.WAITING
            except TimeoutError:
                return False
        return True

    @property
    def is_grammar_ready(self) -> bool:
        return self._check_grammar_completion()

    @property
51
    def grammar(self) -> StructuredOutputGrammar | None:
52
        completed = self._check_grammar_completion()
53
        return (
54
            cast(StructuredOutputGrammar | None, self._grammar) if completed else None
55
        )
56
57

    @grammar.setter
58
    def grammar(
59
        self, grammar: StructuredOutputGrammar | Future[StructuredOutputGrammar]
60
    ) -> None:
61
62
63
64
        self._grammar = grammar

    @functools.cached_property
    def structured_output_key(self) -> StructuredOutputKey:
65
        return get_structured_output_key(self.params)
66
67


68
def get_structured_output_key(params: StructuredOutputsParams) -> StructuredOutputKey:
69
70
71
72
73
    if params.json is not None:
        if not isinstance(params.json, str):
            json_str = json.dumps(params.json)
        else:
            json_str = params.json
74
75
76
77
78
79
        return StructuredOutputOptions.JSON, json_str
    if params.json_object:
        return StructuredOutputOptions.JSON_OBJECT, ""
    if params.regex is not None:
        return StructuredOutputOptions.REGEX, params.regex
    if params.choice is not None:
80
81
        if not isinstance(params.choice, str):
            json_str = json.dumps(params.choice)
82
        else:
83
            json_str = params.choice
84
85
86
87
88
89
        return StructuredOutputOptions.CHOICE, json_str
    if params.grammar is not None:
        return StructuredOutputOptions.GRAMMAR, params.grammar
    if params.structural_tag is not None:
        return StructuredOutputOptions.STRUCTURAL_TAG, params.structural_tag
    raise ValueError("No valid structured output parameter found")