@inference-gateway/sdk 0.4.1-rc.1 → 0.5.0

package/CHANGELOG.md

@@ -2,11 +2,21 @@
 
 All notable changes to this project will be documented in this file.
 
-## [0.4.1-rc.1](https://github.com/inference-gateway/typescript-sdk/compare/v0.4.0...v0.4.1-rc.1) (2025-04-25)
+## [0.5.0](https://github.com/inference-gateway/typescript-sdk/compare/v0.4.1...v0.5.0) (2025-04-26)
+
+### ✨ Features
+
+* Add on reasoning event callback function handling in streaming chat completions ([#7](https://github.com/inference-gateway/typescript-sdk/issues/7)) ([200eb12](https://github.com/inference-gateway/typescript-sdk/commit/200eb12c5890f46d00e5c9cdaaf40a3a306b4315))
+
+### 📦 Miscellaneous
+
+* Add husky pre-commit hook ([#8](https://github.com/inference-gateway/typescript-sdk/issues/8)) ([5ad49a0](https://github.com/inference-gateway/typescript-sdk/commit/5ad49a083e7767a3b7293328c9cead0769324ff6))
+
+## [0.4.1](https://github.com/inference-gateway/typescript-sdk/compare/v0.4.0...v0.4.1) (2025-04-25)
 
 ### ♻️ Improvements
 
-* Enhance tool call handling in streaming chat completions ([f05984a](https://github.com/inference-gateway/typescript-sdk/commit/f05984a3fb8f643ea709e394569a414c7d231b7c))
+* Simplify tool call handling in streaming chat completions ([#6](https://github.com/inference-gateway/typescript-sdk/issues/6)) ([f2d2a5a](https://github.com/inference-gateway/typescript-sdk/commit/f2d2a5a0743b0c1ef4930ef860949f486ee0d5fc))
 
 ## [0.4.0](https://github.com/inference-gateway/typescript-sdk/compare/v0.3.4...v0.4.0) (2025-03-31)
 

package/README.md

@@ -155,11 +155,11 @@ const client = new InferenceGatewayClient({
 try {
   await client.streamChatCompletion(
     {
-      model: 'gpt-4o',
+      model: 'openai/gpt-4o',
       messages: [
         {
           role: MessageRole.User,
-          content: 'What's the weather in San Francisco?',
+          content: "What's the weather in San Francisco?",
         },
       ],
       tools: [
@@ -186,10 +186,14 @@ try {
         console.log('Tool call:', toolCall.function.name);
         console.log('Arguments:', toolCall.function.arguments);
       },
-      onContent: (content) => process.stdout.write(content),
+      onReasoning: (reasoning) => {
+        console.log('Reasoning:', reasoning);
+      },
+      onContent: (content) => {
+        console.log('Content:', content);
+      },
       onFinish: () => console.log('\nStream completed'),
-    },
-    Provider.OpenAI
+    }
   );
 } catch (error) {
   console.error('Error:', error);

package/dist/src/client.d.ts

@@ -2,6 +2,7 @@ import type { Provider, SchemaChatCompletionMessageToolCall, SchemaCreateChatCom
 interface ChatCompletionStreamCallbacks {
     onOpen?: () => void;
     onChunk?: (chunk: SchemaCreateChatCompletionStreamResponse) => void;
+    onReasoning?: (reasoningContent: string) => void;
     onContent?: (content: string) => void;
     onTool?: (toolCall: SchemaChatCompletionMessageToolCall) => void;
     onFinish?: (response: SchemaCreateChatCompletionStreamResponse | null) => void;

package/dist/src/client.js

@@ -162,6 +162,10 @@ class InferenceGatewayClient {
                         try {
                             const chunk = JSON.parse(data);
                             callbacks.onChunk?.(chunk);
+                            const reasoning_content = chunk.choices[0]?.delta?.reasoning_content;
+                            if (reasoning_content !== undefined) {
+                                callbacks.onReasoning?.(reasoning_content);
+                            }
                             const content = chunk.choices[0]?.delta?.content;
                             if (content) {
                                 callbacks.onContent?.(content);

package/dist/tests/client.test.js

@@ -204,6 +204,58 @@ describe('InferenceGatewayClient', () => {
                 }),
             }));
         });
+        it('should handle streaming chat completions reasoning and content', async () => {
+            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();
+            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":"gpt-4o","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"gpt-4o","choices":[{"index":0,"delta":{"content":"","reasoning_content":"This"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"gpt-4o","choices":[{"index":0,"delta":{"content":"","reasoning_content":" is"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"gpt-4o","choices":[{"index":0,"delta":{"content":"","reasoning_content":" a"},"finish_reason":"stop"}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"gpt-4o","choices":[{"index":0,"delta":{"content":"","reasoning_content":" reasoning"},"finish_reason":"stop"}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"gpt-4o","choices":[{"index":0,"delta":{"content":"","reasoning_content":" content"},"finish_reason":"stop"}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"gpt-4o","choices":[{"index":0,"delta":{"content":"Hello"},"finish_reason":null}]}\n\n' +
+                'data: {"id":"chatcmpl-123","object":"chat.completion.chunk","created":1677652288,"model":"gpt-4o","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,
+                }),
+            }));
+        });
         it('should handle tool calls in streaming chat completions', async () => {
             const mockRequest = {
                 model: 'gpt-4o',
@@ -256,8 +308,8 @@ describe('InferenceGatewayClient', () => {
                 type: 'function',
                 function: {
                     name: 'get_weather',
-                    arguments: '{"location":"San Francisco, CA"}'
-                }
+                    arguments: '{"location":"San Francisco, CA"}',
+                },
             });
             expect(callbacks.onFinish).toHaveBeenCalledTimes(1);
         });

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@inference-gateway/sdk",
-  "version": "0.4.1-rc.1",
+  "version": "0.5.0",
   "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",
@@ -49,7 +49,8 @@
     "build": "tsc",
     "test": "jest",
     "lint": "eslint src/**/*.ts",
-    "prepare": "npm run build"
+    "format": "prettier --write \"src/**/*.ts\" \"tests/**/*.ts\"",
+    "prepare": "husky"
   },
   "devDependencies": {
     "@eslint/js": "^9.18.0",
@@ -59,7 +60,9 @@
     "@typescript-eslint/parser": "^8.21.0",
     "eslint": "^9.18.0",
     "eslint-plugin-prettier": "^5.2.3",
+    "husky": "^9.1.7",
     "jest": "^29.7.0",
+    "prettier": "^3.5.3",
     "ts-jest": "^29.2.5",
     "typescript": "^5.7.3"
   },