@inference-gateway/sdk 0.6.0 → 0.6.2

package/CHANGELOG.md

@@ -2,6 +2,18 @@
 
 All notable changes to this project will be documented in this file.
 
+## [0.6.2](https://github.com/inference-gateway/typescript-sdk/compare/v0.6.1...v0.6.2) (2025-04-30)
+
+### ♻️ Improvements
+
+* Process also groq reasoning models properly ([#12](https://github.com/inference-gateway/typescript-sdk/issues/12)) ([51ce3bb](https://github.com/inference-gateway/typescript-sdk/commit/51ce3bbbbdf03947bb7928e8edc413b977ea092a))
+
+## [0.6.1](https://github.com/inference-gateway/typescript-sdk/compare/v0.6.0...v0.6.1) (2025-04-28)
+
+### ♻️ Improvements
+
+*  Remove redundant request option ([#11](https://github.com/inference-gateway/typescript-sdk/issues/11)) ([82e34e2](https://github.com/inference-gateway/typescript-sdk/commit/82e34e2ee9782fd224945bff1bd4daf2859a4f79))
+
 ## [0.6.0](https://github.com/inference-gateway/typescript-sdk/compare/v0.5.1...v0.6.0) (2025-04-28)
 
 ### ✨ Features

package/dist/src/client.d.ts

@@ -40,11 +40,16 @@ export declare class InferenceGatewayClient {
     /**
      * Creates a chat completion.
      */
-    createChatCompletion(request: SchemaCreateChatCompletionRequest, provider?: Provider): Promise<SchemaCreateChatCompletionResponse>;
+    createChatCompletion(request: Omit<SchemaCreateChatCompletionRequest, 'stream'>, provider?: Provider): Promise<SchemaCreateChatCompletionResponse>;
     /**
      * Creates a streaming chat completion.
+     * This method always sets stream=true internally, so there's no need to specify it in the request.
+     *
+     * @param request - Chat completion request (must include at least model and messages)
+     * @param callbacks - Callbacks for handling streaming events
+     * @param provider - Optional provider to use for this request
      */
-    streamChatCompletion(request: SchemaCreateChatCompletionRequest, callbacks: ChatCompletionStreamCallbacks, provider?: Provider): Promise<void>;
+    streamChatCompletion(request: Omit<SchemaCreateChatCompletionRequest, 'stream' | 'stream_options'>, callbacks: ChatCompletionStreamCallbacks, provider?: Provider): Promise<void>;
     /**
      * Proxy a request to a specific provider.
      */

package/dist/src/client.js

@@ -87,11 +87,16 @@ class InferenceGatewayClient {
         }
         return this.request('/chat/completions', {
             method: 'POST',
-            body: JSON.stringify(request),
+            body: JSON.stringify({ ...request, stream: false }),
         }, query);
     }
     /**
      * Creates a streaming chat completion.
+     * This method always sets stream=true internally, so there's no need to specify it in the request.
+     *
+     * @param request - Chat completion request (must include at least model and messages)
+     * @param callbacks - Callbacks for handling streaming events
+     * @param provider - Optional provider to use for this request
      */
     async streamChatCompletion(request, callbacks, provider) {
         const query = {};
@@ -120,6 +125,9 @@ class InferenceGatewayClient {
                 body: JSON.stringify({
                     ...request,
                     stream: true,
+                    stream_options: {
+                        include_usage: true,
+                    },
                 }),
                 signal: controller.signal,
             });
@@ -169,6 +177,10 @@ class InferenceGatewayClient {
                             if (reasoning_content !== undefined) {
                                 callbacks.onReasoning?.(reasoning_content);
                             }
+                            const reasoning = chunk.choices[0]?.delta?.reasoning;
+                            if (reasoning !== undefined) {
+                                callbacks.onReasoning?.(reasoning);
+                            }
                             const content = chunk.choices[0]?.delta?.content;
                             if (content) {
                                 callbacks.onContent?.(content);

package/dist/src/types/generated/index.d.ts

@@ -180,8 +180,8 @@ export interface components {
             retry?: number;
         };
         Endpoints: {
-            models?: string;
-            chat?: string;
+            models: string;
+            chat: string;
         };
         Error: {
             error?: string;
@@ -197,17 +197,19 @@ export interface components {
             content: string;
             tool_calls?: components['schemas']['ChatCompletionMessageToolCall'][];
             tool_call_id?: string;
-            reasoning?: string;
+            /** @description The reasoning content of the chunk message. */
             reasoning_content?: string;
+            /** @description The reasoning of the chunk message. Same as reasoning_content. */
+            reasoning?: string;
         };
         /** @description Common model information */
         Model: {
-            id?: string;
-            object?: string;
+            id: string;
+            object: string;
             /** Format: int64 */
-            created?: number;
-            owned_by?: string;
-            served_by?: components['schemas']['Provider'];
+            created: number;
+            owned_by: string;
+            served_by: components['schemas']['Provider'];
         };
         /** @description Response structure for listing models */
         ListModelsResponse: {
@@ -266,11 +268,8 @@ export interface components {
         /** @description Options for streaming response. Only set this when you set `stream: true`.
          *      */
         ChatCompletionStreamOptions: {
-            /**
-             * @description If set, an additional chunk will be streamed before the `data: [DONE]` message. The `usage` field on this chunk shows the token usage statistics for the entire request, and the `choices` field will always be an empty array. All other chunks will also include a `usage` field, but with a null value.
-             *
-             * @default true
-             */
+            /** @description If set, an additional chunk will be streamed before the `data: [DONE]` message. The `usage` field on this chunk shows the token usage statistics for the entire request, and the `choices` field will always be an empty array. All other chunks will also include a `usage` field, but with a null value.
+             *      */
             include_usage: boolean;
         };
         CreateChatCompletionRequest: {
@@ -292,6 +291,10 @@ export interface components {
             /** @description A list of tools the model may call. Currently, only functions are supported as a tool. Use this to provide a list of functions the model may generate JSON inputs for. A max of 128 functions are supported.
              *      */
             tools?: components['schemas']['ChatCompletionTool'][];
+            /** @description The format of the reasoning content. Can be `raw` or `parsed`.
+             *     When specified as raw some reasoning models will output <think /> tags. When specified as parsed the model will output the reasoning under  `reasoning` or `reasoning_content` attribute.
+             *      */
+            reasoning_format?: string;
         };
         /** @description The function that the model called. */
         ChatCompletionMessageToolCallFunction: {
@@ -350,11 +353,13 @@ export interface components {
         /** @description A chat completion delta generated by streamed model responses. */
         ChatCompletionStreamResponseDelta: {
             /** @description The contents of the chunk message. */
-            content?: string;
+            content: string;
             /** @description The reasoning content of the chunk message. */
             reasoning_content?: string;
+            /** @description The reasoning of the chunk message. Same as reasoning_content. */
+            reasoning?: string;
             tool_calls?: components['schemas']['ChatCompletionMessageToolCallChunk'][];
-            role?: components['schemas']['MessageRole'];
+            role: components['schemas']['MessageRole'];
             /** @description The refusal message generated by the model. */
             refusal?: string;
         };
@@ -418,6 +423,10 @@ export interface components {
             /** @description The object type, which is always `chat.completion.chunk`. */
             object: string;
             usage?: components['schemas']['CompletionUsage'];
+            /** @description The format of the reasoning content. Can be `raw` or `parsed`.
+             *     When specified as raw some reasoning models will output <think /> tags. When specified as parsed the model will output the reasoning under reasoning_content.
+             *      */
+            reasoning_format?: string;
         };
         Config: unknown;
     };

package/dist/tests/client.test.js

@@ -57,6 +57,7 @@ describe('InferenceGatewayClient', () => {
                         object: 'model',
                         created: 1686935002,
                         owned_by: 'openai',
+                        served_by: generated_1.Provider.openai,
                     },
                 ],
             };
@@ -89,7 +90,6 @@ describe('InferenceGatewayClient', () => {
                     { role: generated_1.MessageRole.system, content: 'You are a helpful assistant' },
                     { role: generated_1.MessageRole.user, content: 'Hello' },
                 ],
-                stream: false,
             };
             const mockResponse = {
                 id: 'chatcmpl-123',
@@ -120,14 +120,13 @@ describe('InferenceGatewayClient', () => {
             expect(result).toEqual(mockResponse);
             expect(mockFetch).toHaveBeenCalledWith('http://localhost:8080/v1/chat/completions', expect.objectContaining({
                 method: 'POST',
-                body: JSON.stringify(mockRequest),
+                body: JSON.stringify({ ...mockRequest, stream: false }),
             }));
         });
         it('should create a chat completion with a specific provider', async () => {
             const mockRequest = {
                 model: 'claude-3-opus-20240229',
                 messages: [{ role: generated_1.MessageRole.user, content: 'Hello' }],
-                stream: false,
             };
             const mockResponse = {
                 id: 'chatcmpl-456',
@@ -158,7 +157,7 @@ describe('InferenceGatewayClient', () => {
             expect(result).toEqual(mockResponse);
             expect(mockFetch).toHaveBeenCalledWith('http://localhost:8080/v1/chat/completions?provider=anthropic', expect.objectContaining({
                 method: 'POST',
-                body: JSON.stringify(mockRequest),
+                body: JSON.stringify({ ...mockRequest, stream: false }),
             }));
         });
     });
@@ -167,7 +166,6 @@ describe('InferenceGatewayClient', () => {
             const mockRequest = {
                 model: 'gpt-4o',
                 messages: [{ role: generated_1.MessageRole.user, content: 'Hello' }],
-                stream: true,
             };
             const mockStream = new web_1.TransformStream();
             const writer = mockStream.writable.getWriter();
@@ -201,6 +199,9 @@ describe('InferenceGatewayClient', () => {
                 body: JSON.stringify({
                     ...mockRequest,
                     stream: true,
+                    stream_options: {
+                        include_usage: true,
+                    },
                 }),
             }));
         });
@@ -208,7 +209,6 @@ describe('InferenceGatewayClient', () => {
             const mockRequest = {
                 model: 'gpt-4o',
                 messages: [{ role: generated_1.MessageRole.user, content: 'Hello' }],
-                stream: true,
             };
             const mockStream = new web_1.TransformStream();
             const writer = mockStream.writable.getWriter();
@@ -253,6 +253,9 @@ describe('InferenceGatewayClient', () => {
                 body: JSON.stringify({
                     ...mockRequest,
                     stream: true,
+                    stream_options: {
+                        include_usage: true,
+                    },
                 }),
             }));
         });
@@ -274,7 +277,6 @@ describe('InferenceGatewayClient', () => {
                         },
                     },
                 ],
-                stream: true,
             };
             const mockStream = new web_1.TransformStream();
             const writer = mockStream.writable.getWriter();
@@ -312,12 +314,21 @@ describe('InferenceGatewayClient', () => {
                 },
             });
             expect(callbacks.onFinish).toHaveBeenCalledTimes(1);
+            expect(mockFetch).toHaveBeenCalledWith('http://localhost:8080/v1/chat/completions', expect.objectContaining({
+                method: 'POST',
+                body: JSON.stringify({
+                    ...mockRequest,
+                    stream: true,
+                    stream_options: {
+                        include_usage: true,
+                    },
+                }),
+            }));
         });
         it('should handle errors in streaming chat completions', async () => {
             const mockRequest = {
                 model: 'gpt-4o',
                 messages: [{ role: generated_1.MessageRole.user, content: 'Hello' }],
-                stream: true,
             };
             mockFetch.mockResolvedValueOnce({
                 ok: false,
@@ -334,10 +345,6 @@ describe('InferenceGatewayClient', () => {
             const mockRequest = {
                 model: 'gpt-4o',
                 messages: [{ role: generated_1.MessageRole.user, content: 'Hello' }],
-                stream: true,
-                stream_options: {
-                    include_usage: true,
-                },
             };
             const mockStream = new web_1.TransformStream();
             const writer = mockStream.writable.getWriter();
@@ -379,6 +386,167 @@ describe('InferenceGatewayClient', () => {
                 body: JSON.stringify({
                     ...mockRequest,
                     stream: true,
+                    stream_options: {
+                        include_usage: true,
+                    },
+                }),
+            }));
+        });
+        it('should handle streaming chat completions with reasoning field', async () => {
+            const mockRequest = {
+                model: 'groq/deepseek-distilled-llama-3.1-70b',
+                messages: [{ role: generated_1.MessageRole.user, content: 'Hello' }],
+            };
+            const mockStream = new web_1.TransformStream();
+            const writer = mockStream.writable.getWriter();
+            const encoder = new node_util_1.TextEncoder();
+            mockFetch.mockResolvedValueOnce({
+                ok: true,
+                body: mockStream.readable,
+            });
+            const callbacks = {
+                onOpen: jest.fn(),
+                onChunk: jest.fn(),
+                onReasoning: jest.fn(),
+                onContent: jest.fn(),
+                onFinish: jest.fn(),
+            };
+            const streamPromise = client.streamChatCompletion(mockRequest, callbacks);
+            await writer.write(encoder.encode('data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"groq/deepseek-distilled-llama-3.1-70b","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"groq/deepseek-distilled-llama-3.1-70b","choices":[{"index":0,"delta":{"content":"","reasoning":"Let me"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"groq/deepseek-distilled-llama-3.1-70b","choices":[{"index":0,"delta":{"content":"","reasoning":" think"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"groq/deepseek-distilled-llama-3.1-70b","choices":[{"index":0,"delta":{"content":"","reasoning":" about"},"finish_reason":"stop"}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"groq/deepseek-distilled-llama-3.1-70b","choices":[{"index":0,"delta":{"content":"","reasoning":" this"},"finish_reason":"stop"}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"groq/deepseek-distilled-llama-3.1-70b","choices":[{"index":0,"delta":{"content":"Hello"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"groq/deepseek-distilled-llama-3.1-70b","choices":[{"index":0,"delta":{"content":"!"},"finish_reason":null}]}\n\n' +
+                'data: [DONE]\n\n'));
+            await writer.close();
+            await streamPromise;
+            expect(callbacks.onOpen).toHaveBeenCalledTimes(1);
+            expect(callbacks.onChunk).toHaveBeenCalledTimes(7);
+            expect(callbacks.onReasoning).toHaveBeenCalledTimes(4);
+            expect(callbacks.onReasoning).toHaveBeenCalledWith('Let me');
+            expect(callbacks.onReasoning).toHaveBeenCalledWith(' think');
+            expect(callbacks.onReasoning).toHaveBeenCalledWith(' about');
+            expect(callbacks.onReasoning).toHaveBeenCalledWith(' this');
+            expect(callbacks.onContent).toHaveBeenCalledTimes(2);
+            expect(callbacks.onContent).toHaveBeenCalledWith('Hello');
+            expect(callbacks.onContent).toHaveBeenCalledWith('!');
+            expect(callbacks.onFinish).toHaveBeenCalledTimes(1);
+            expect(mockFetch).toHaveBeenCalledWith('http://localhost:8080/v1/chat/completions', expect.objectContaining({
+                method: 'POST',
+                body: JSON.stringify({
+                    ...mockRequest,
+                    stream: true,
+                    stream_options: {
+                        include_usage: true,
+                    },
+                }),
+            }));
+        });
+        it('should handle streaming chat completions with reasoning_content (DeepSeek)', async () => {
+            const mockRequest = {
+                model: 'deepseek/deepseek-reasoner',
+                messages: [{ role: generated_1.MessageRole.user, content: 'Hello' }],
+            };
+            const mockStream = new web_1.TransformStream();
+            const writer = mockStream.writable.getWriter();
+            const encoder = new node_util_1.TextEncoder();
+            mockFetch.mockResolvedValueOnce({
+                ok: true,
+                body: mockStream.readable,
+            });
+            const callbacks = {
+                onOpen: jest.fn(),
+                onChunk: jest.fn(),
+                onReasoning: jest.fn(),
+                onContent: jest.fn(),
+                onFinish: jest.fn(),
+            };
+            const streamPromise = client.streamChatCompletion(mockRequest, callbacks);
+            await writer.write(encoder.encode('data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"deepseek/deepseek-reasoner","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"deepseek/deepseek-reasoner","choices":[{"index":0,"delta":{"content":"","reasoning_content":"This"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"deepseek/deepseek-reasoner","choices":[{"index":0,"delta":{"content":"","reasoning_content":" is"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"deepseek/deepseek-reasoner","choices":[{"index":0,"delta":{"content":"","reasoning_content":" a"},"finish_reason":"stop"}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"deepseek/deepseek-reasoner","choices":[{"index":0,"delta":{"content":"","reasoning_content":" reasoning"},"finish_reason":"stop"}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"deepseek/deepseek-reasoner","choices":[{"index":0,"delta":{"content":"","reasoning_content":" content"},"finish_reason":"stop"}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"deepseek/deepseek-reasoner","choices":[{"index":0,"delta":{"content":"Hello"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"deepseek/deepseek-reasoner","choices":[{"index":0,"delta":{"content":"!"},"finish_reason":null}]}\n\n' +
+                'data: [DONE]\n\n'));
+            await writer.close();
+            await streamPromise;
+            expect(callbacks.onOpen).toHaveBeenCalledTimes(1);
+            expect(callbacks.onChunk).toHaveBeenCalledTimes(8);
+            expect(callbacks.onReasoning).toHaveBeenCalledTimes(5);
+            expect(callbacks.onReasoning).toHaveBeenCalledWith('This');
+            expect(callbacks.onReasoning).toHaveBeenCalledWith(' is');
+            expect(callbacks.onReasoning).toHaveBeenCalledWith(' a');
+            expect(callbacks.onReasoning).toHaveBeenCalledWith(' reasoning');
+            expect(callbacks.onReasoning).toHaveBeenCalledWith(' content');
+            expect(callbacks.onContent).toHaveBeenCalledTimes(2);
+            expect(callbacks.onContent).toHaveBeenCalledWith('Hello');
+            expect(callbacks.onContent).toHaveBeenCalledWith('!');
+            expect(callbacks.onFinish).toHaveBeenCalledTimes(1);
+            expect(mockFetch).toHaveBeenCalledWith('http://localhost:8080/v1/chat/completions', expect.objectContaining({
+                method: 'POST',
+                body: JSON.stringify({
+                    ...mockRequest,
+                    stream: true,
+                    stream_options: {
+                        include_usage: true,
+                    },
+                }),
+            }));
+        });
+        it('should handle streaming chat completions with reasoning field (Groq)', async () => {
+            const mockRequest = {
+                model: 'llama-3.1-70b-versatile',
+                messages: [{ role: generated_1.MessageRole.user, content: 'Hello' }],
+            };
+            const mockStream = new web_1.TransformStream();
+            const writer = mockStream.writable.getWriter();
+            const encoder = new node_util_1.TextEncoder();
+            mockFetch.mockResolvedValueOnce({
+                ok: true,
+                body: mockStream.readable,
+            });
+            const callbacks = {
+                onOpen: jest.fn(),
+                onChunk: jest.fn(),
+                onReasoning: jest.fn(),
+                onContent: jest.fn(),
+                onFinish: jest.fn(),
+            };
+            const streamPromise = client.streamChatCompletion(mockRequest, callbacks);
+            await writer.write(encoder.encode('data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"llama-3.1-70b-versatile","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"llama-3.1-70b-versatile","choices":[{"index":0,"delta":{"content":"","reasoning":"Let me"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"llama-3.1-70b-versatile","choices":[{"index":0,"delta":{"content":"","reasoning":" think"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"llama-3.1-70b-versatile","choices":[{"index":0,"delta":{"content":"","reasoning":" about"},"finish_reason":"stop"}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"llama-3.1-70b-versatile","choices":[{"index":0,"delta":{"content":"","reasoning":" this"},"finish_reason":"stop"}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"llama-3.1-70b-versatile","choices":[{"index":0,"delta":{"content":"Hello"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"llama-3.1-70b-versatile","choices":[{"index":0,"delta":{"content":"!"},"finish_reason":null}]}\n\n' +
+                'data: [DONE]\n\n'));
+            await writer.close();
+            await streamPromise;
+            expect(callbacks.onOpen).toHaveBeenCalledTimes(1);
+            expect(callbacks.onChunk).toHaveBeenCalledTimes(7);
+            expect(callbacks.onReasoning).toHaveBeenCalledTimes(4);
+            expect(callbacks.onReasoning).toHaveBeenCalledWith('Let me');
+            expect(callbacks.onReasoning).toHaveBeenCalledWith(' think');
+            expect(callbacks.onReasoning).toHaveBeenCalledWith(' about');
+            expect(callbacks.onReasoning).toHaveBeenCalledWith(' this');
+            expect(callbacks.onContent).toHaveBeenCalledTimes(2);
+            expect(callbacks.onContent).toHaveBeenCalledWith('Hello');
+            expect(callbacks.onContent).toHaveBeenCalledWith('!');
+            expect(callbacks.onFinish).toHaveBeenCalledTimes(1);
+            expect(mockFetch).toHaveBeenCalledWith('http://localhost:8080/v1/chat/completions', expect.objectContaining({
+                method: 'POST',
+                body: JSON.stringify({
+                    ...mockRequest,
+                    stream: true,
+                    stream_options: {
+                        include_usage: true,
+                    },
                 }),
             }));
         });

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@inference-gateway/sdk",
-  "version": "0.6.0",
+  "version": "0.6.2",
   "description": "An SDK written in Typescript for the [Inference Gateway](https://github.com/inference-gateway/inference-gateway).",
   "main": "dist/src/index.js",
   "types": "dist/src/index.d.ts",