dynamo_llm/
types.rs

1// SPDX-FileCopyrightText: Copyright (c) 2024-2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
2// SPDX-License-Identifier: Apache-2.0
3
4use crate::protocols;
5
6pub use protocols::{Annotated, TokenIdType};
7
8pub mod openai {
9    use super::*;
10    use dynamo_runtime::pipeline::{ServerStreamingEngine, UnaryEngine};
11
12    pub mod completions {
13        use super::*;
14
15        pub use protocols::openai::completions::{
16            NvCreateCompletionRequest, NvCreateCompletionResponse,
17        };
18
19        /// A [`UnaryEngine`] implementation for the OpenAI Completions API
20        pub type OpenAICompletionsUnaryEngine =
21            UnaryEngine<NvCreateCompletionRequest, NvCreateCompletionResponse>;
22
23        /// A [`ServerStreamingEngine`] implementation for the OpenAI Completions API
24        pub type OpenAICompletionsStreamingEngine =
25            ServerStreamingEngine<NvCreateCompletionRequest, Annotated<NvCreateCompletionResponse>>;
26    }
27
28    pub mod chat_completions {
29        use super::*;
30
31        pub use protocols::openai::chat_completions::{
32            NvCreateChatCompletionRequest, NvCreateChatCompletionResponse,
33            NvCreateChatCompletionStreamResponse,
34        };
35
36        /// A [`UnaryEngine`] implementation for the OpenAI Chat Completions API
37        pub type OpenAIChatCompletionsUnaryEngine =
38            UnaryEngine<NvCreateChatCompletionRequest, NvCreateChatCompletionResponse>;
39
40        /// A [`ServerStreamingEngine`] implementation for the OpenAI Chat Completions API
41        pub type OpenAIChatCompletionsStreamingEngine = ServerStreamingEngine<
42            NvCreateChatCompletionRequest,
43            Annotated<NvCreateChatCompletionStreamResponse>,
44        >;
45    }
46
47    pub mod embeddings {
48        use super::*;
49
50        pub use protocols::openai::embeddings::{
51            NvCreateEmbeddingRequest, NvCreateEmbeddingResponse,
52        };
53
54        /// A [`UnaryEngine`] implementation for the OpenAI Embeddings API
55        pub type OpenAIEmbeddingsUnaryEngine =
56            UnaryEngine<NvCreateEmbeddingRequest, NvCreateEmbeddingResponse>;
57
58        /// A [`ServerStreamingEngine`] implementation for the OpenAI Embeddings API
59        pub type OpenAIEmbeddingsStreamingEngine =
60            ServerStreamingEngine<NvCreateEmbeddingRequest, Annotated<NvCreateEmbeddingResponse>>;
61    }
62}
63
64pub mod generic {
65    use super::*;
66    use dynamo_runtime::pipeline::{ServerStreamingEngine, UnaryEngine};
67
68    pub mod tensor {
69        use super::*;
70
71        pub use protocols::tensor::{NvCreateTensorRequest, NvCreateTensorResponse};
72
73        /// A [`UnaryEngine`] implementation for the generic Tensor API
74        pub type TensorUnaryEngine = UnaryEngine<NvCreateTensorRequest, NvCreateTensorResponse>;
75
76        /// A [`ServerStreamingEngine`] implementation for the generic Tensor API
77        pub type TensorStreamingEngine =
78            ServerStreamingEngine<NvCreateTensorRequest, Annotated<NvCreateTensorResponse>>;
79    }
80}
dynamo_llm/types.rs

dynamo_llm/
types.rs
