| 1 | # SPDX-License-Identifier: Apache-2.0 |
| 2 | # SPDX-FileCopyrightText: Copyright contributors to the vLLM project |
| 3 | |
| 4 | from typing import Protocol, TypeAlias |
| 5 | |
| 6 | from vllm.config import ModelConfig |
| 7 | from vllm.entrypoints.chat_utils import ChatTemplateContentFormatOption |
| 8 | from vllm.entrypoints.openai.chat_completion.protocol import ( |
| 9 | BatchChatCompletionRequest, |
| 10 | ChatCompletionRequest, |
| 11 | ChatCompletionResponse, |
| 12 | ) |
| 13 | from vllm.entrypoints.openai.completion.protocol import ( |
| 14 | CompletionRequest, |
| 15 | CompletionResponse, |
| 16 | ) |
| 17 | from vllm.entrypoints.openai.responses.protocol import ResponsesRequest |
| 18 | from vllm.entrypoints.scale_out.token_in_token_out.protocol import ( |
| 19 | DerenderChatRequest, |
| 20 | DerenderChatStreamRequest, |
| 21 | DerenderCompletionRequest, |
| 22 | DerenderCompletionStreamRequest, |
| 23 | GenerateRequest, |
| 24 | GenerateResponse, |
| 25 | ) |
| 26 | from vllm.entrypoints.serve.tokenize.protocol import ( |
| 27 | DetokenizeRequest, |
| 28 | TokenizeChatRequest, |
| 29 | TokenizeCompletionRequest, |
| 30 | TokenizeResponse, |
| 31 | ) |
| 32 | from vllm.entrypoints.speech_to_text.transcription.protocol import ( |
| 33 | TranscriptionRequest, |
| 34 | TranscriptionResponse, |
| 35 | ) |
| 36 | from vllm.entrypoints.speech_to_text.translation.protocol import TranslationRequest |
| 37 | from vllm.renderers import ChatParams, TokenizeParams |
| 38 | |
| 39 | |
| 40 | class RendererRequest(Protocol): |
| 41 | def build_tok_params(self, model_config: ModelConfig) -> TokenizeParams: |
| 42 | raise NotImplementedError |
| 43 | |
| 44 | |
| 45 | class RendererChatRequest(RendererRequest, Protocol): |
| 46 | def build_chat_params( |
| 47 | self, |
| 48 | default_template: str | None, |
| 49 | default_template_content_format: ChatTemplateContentFormatOption, |
| 50 | ) -> ChatParams: |
| 51 | raise NotImplementedError |
| 52 | |
| 53 | |
| 54 | CompletionLikeRequest: TypeAlias = ( |
| 55 | CompletionRequest |
| 56 | | TokenizeCompletionRequest |
| 57 | | DetokenizeRequest |
| 58 | | DerenderCompletionRequest |
| 59 | | DerenderCompletionStreamRequest |
| 60 | ) |
| 61 | |
| 62 | ChatLikeRequest: TypeAlias = ( |
| 63 | ChatCompletionRequest |
| 64 | | BatchChatCompletionRequest |
| 65 | | TokenizeChatRequest |
| 66 | | DerenderChatRequest |
| 67 | | DerenderChatStreamRequest |
| 68 | ) |
| 69 | |
| 70 | SpeechToTextRequest: TypeAlias = TranscriptionRequest | TranslationRequest |
| 71 | |
| 72 | AnyRequest: TypeAlias = ( |
| 73 | CompletionLikeRequest |
| 74 | | ChatLikeRequest |
| 75 | | SpeechToTextRequest |
| 76 | | ResponsesRequest |
| 77 | | GenerateRequest |
| 78 | ) |
| 79 | |
| 80 | AnyResponse: TypeAlias = ( |
| 81 | CompletionResponse |
| 82 | | ChatCompletionResponse |
| 83 | | TranscriptionResponse |
| 84 | | TokenizeResponse |
| 85 | | GenerateResponse |
| 86 | ) |