ai-world-sdk 1.0.8 → 1.0.9

package/README.md

@@ -247,6 +247,34 @@ const result3 = await aihubmixClient.generate({
   number_of_images: 1,
   response_modalities: ['IMAGE'],
 });
+
+// 图像编辑（使用文本提示编辑图片）
+const editResult = await client.edit({
+  prompt: 'Add a small wizard hat on the cat\'s head',
+  image: 'data:image/png;base64,iVBORw0KGgo...', // base64 编码的图片数据或 data URL
+  model: 'gemini-2.5-flash-image',
+  aspect_ratio: '1:1',
+  response_modalities: ['IMAGE'],
+});
+
+// 多轮图片修改（迭代式优化图片）
+// 第一轮：创建初始图片
+const firstResponse = await client.chat({
+  message: 'Create a vibrant infographic about photosynthesis',
+  model: 'gemini-3-pro-image-preview',
+  aspect_ratio: '16:9',
+  response_modalities: ['TEXT', 'IMAGE'],
+});
+
+// 后续轮次：修改图片（使用返回的 chat_id）
+const secondResponse = await client.chat({
+  chat_id: firstResponse.chat_id,
+  message: 'Update this infographic to be in Spanish',
+  model: 'gemini-3-pro-image-preview',
+  aspect_ratio: '16:9',
+  image_size: '2K',
+  response_modalities: ['TEXT', 'IMAGE'],
+});
 ```
 
 **参数说明：**
@@ -276,6 +304,17 @@ const result3 = await aihubmixClient.generate({
 | `gemini-2.5-flash-image` | 1024px | 快速、高效、成本低 | 日常使用、批量生成 |
 | `gemini-3-pro-image-preview` | 1K/2K/4K | 专业级、高分辨率、高级功能 | 专业设计、高分辨率需求 |
 
+**图像编辑和多轮修改：**
+
+- **`edit()`** - 图像编辑：使用文本提示编辑图片，支持添加、移除或修改元素
+  - 需要提供输入图片（base64 编码或 data URL）
+  - 适用于单次编辑操作
+
+- **`chat()`** - 多轮图片修改：通过对话迭代式优化图片
+  - 首次调用创建新的聊天会话，返回 `chat_id`
+  - 后续调用使用 `chat_id` 继续对话
+  - 推荐使用 `gemini-3-pro-image-preview` 模型进行多轮编辑
+
 ### 视频生成
 
 #### VideoGenerationClient

package/dist/__tests__/example.test.js

@@ -610,6 +610,92 @@ describe("Langchain SDK Tests", () => {
             console.log("图像描述:", result.text);
         }
     }, 120000);
+    test("GeminiImageGenerationClient - 图像编辑", async () => {
+        const imageClient = new index_1.GeminiImageGenerationClient({});
+        // 首先生成一张基础图片
+        const generateResult = await imageClient.generate({
+            prompt: 'A simple red apple on a white background',
+            model: 'gemini-2.5-flash-image',
+            aspect_ratio: '1:1',
+            response_modalities: ['IMAGE'],
+        });
+        expect(generateResult).toBeDefined();
+        expect(generateResult.data).toBeDefined();
+        expect(generateResult.data.length).toBeGreaterThan(0);
+        // 获取第一张图片的 base64 数据
+        const firstImage = generateResult.data[0];
+        expect(firstImage).toBeDefined();
+        // 提取 base64 数据（优先使用 b64_json，否则从 text 中提取）
+        let base64Image;
+        if (firstImage.b64_json) {
+            base64Image = firstImage.b64_json;
+        }
+        else if (firstImage.text && firstImage.text.startsWith('data:')) {
+            // 从 data URL 中提取 base64 部分
+            base64Image = firstImage.text.split(',')[1];
+        }
+        else {
+            throw new Error('无法获取图片的 base64 数据');
+        }
+        expect(base64Image).toBeDefined();
+        // 使用编辑功能添加元素
+        const editResult = await imageClient.edit({
+            prompt: 'Add a small green leaf on top of the apple',
+            image: base64Image, // 直接使用 base64 字符串
+            model: 'gemini-2.5-flash-image',
+            aspect_ratio: '1:1',
+            response_modalities: ['IMAGE'],
+        });
+        expect(editResult).toBeDefined();
+        expect(editResult.created).toBeDefined();
+        expect(editResult.data).toBeDefined();
+        expect(Array.isArray(editResult.data)).toBe(true);
+        expect(editResult.data.length).toBeGreaterThan(0);
+        editResult.data.forEach((item) => {
+            expect(item).toBeDefined();
+            expect(item.url || item.b64_json).toBeDefined();
+        });
+        console.log("✅ GeminiImageGenerationClient 图像编辑测试成功");
+        console.log(`编辑后图像数量: ${editResult.data.length}`);
+    }, 180000);
+    test("GeminiImageGenerationClient - 多轮图片修改", async () => {
+        const imageClient = new index_1.GeminiImageGenerationClient({});
+        // 第一轮：创建初始图片
+        const firstResponse = await imageClient.chat({
+            message: 'Create a vibrant infographic that explains photosynthesis as if it were a recipe for a plant\'s favorite food. Show the "ingredients" (sunlight, water, CO2) and the "finished dish" (sugar/energy). The style should be like a page from a colorful kids\' cookbook, suitable for a 4th grader.',
+            model: 'gemini-3-pro-image-preview',
+            aspect_ratio: '16:9',
+            response_modalities: ['TEXT', 'IMAGE'],
+        });
+        expect(firstResponse).toBeDefined();
+        expect(firstResponse.chat_id).toBeDefined();
+        expect(typeof firstResponse.chat_id).toBe("string");
+        expect(firstResponse.data).toBeDefined();
+        expect(Array.isArray(firstResponse.data)).toBe(true);
+        expect(firstResponse.data.length).toBeGreaterThan(0);
+        console.log("✅ 第一轮图片生成成功");
+        console.log(`Chat ID: ${firstResponse.chat_id}`);
+        console.log(`生成图像数量: ${firstResponse.data.length}`);
+        // 第二轮：修改图片（将语言改为西班牙语）
+        const secondResponse = await imageClient.chat({
+            chat_id: firstResponse.chat_id,
+            message: 'Update this infographic to be in Spanish. Do not change any other elements of the image.',
+            model: 'gemini-3-pro-image-preview',
+            aspect_ratio: '16:9',
+            image_size: '2K',
+            response_modalities: ['TEXT', 'IMAGE'],
+        });
+        expect(secondResponse).toBeDefined();
+        expect(secondResponse.chat_id).toBe(firstResponse.chat_id); // chat_id 应该保持一致
+        expect(secondResponse.data).toBeDefined();
+        expect(Array.isArray(secondResponse.data)).toBe(true);
+        expect(secondResponse.data.length).toBeGreaterThan(0);
+        console.log("✅ 第二轮图片修改成功");
+        console.log(`修改后图像数量: ${secondResponse.data.length}`);
+        if (secondResponse.text) {
+            console.log("文本响应:", secondResponse.text);
+        }
+    }, 240000);
     test("DoubaoImageGenerationClient - quality 参数测试", async () => {
         const imageClient = new index_1.DoubaoImageGenerationClient({});
         const qualities = ["standard", "hd"];

package/dist/gemini-image-generation.d.ts

@@ -28,6 +28,32 @@ export interface GeminiImageGenerationResponse {
     data: GeminiImageData[];
     text?: string;
 }
+export interface GeminiImageEditRequest {
+    prompt: string;
+    image: string;
+    model?: string;
+    provider?: "aihubmix" | "gemini";
+    aspect_ratio?: "1:1" | "2:3" | "3:2" | "3:4" | "4:3" | "4:5" | "5:4" | "9:16" | "16:9" | "21:9";
+    image_size?: "1K" | "2K" | "4K";
+    response_modalities?: ("TEXT" | "IMAGE")[];
+    user?: string;
+}
+export interface GeminiImageChatRequest {
+    message: string;
+    chat_id?: string;
+    model?: string;
+    provider?: "aihubmix" | "gemini";
+    aspect_ratio?: "1:1" | "2:3" | "3:2" | "3:4" | "4:3" | "4:5" | "5:4" | "9:16" | "16:9" | "21:9";
+    image_size?: "1K" | "2K" | "4K";
+    response_modalities?: ("TEXT" | "IMAGE")[];
+    user?: string;
+}
+export interface GeminiImageChatResponse {
+    chat_id: string;
+    created: number;
+    data: GeminiImageData[];
+    text?: string;
+}
 export declare class GeminiImageGenerationClient {
     private headers;
     private provider;
@@ -43,4 +69,45 @@ export declare class GeminiImageGenerationClient {
      * - number_of_images: 生成图片数量（1-4）
      */
     generate(request: GeminiImageGenerationRequest): Promise<GeminiImageGenerationResponse>;
+    /**
+     * Edit images
+     * 编辑图像
+     *
+     * 使用文本提示编辑图片，支持添加、移除或修改元素
+     *
+     * 支持的参数：
+     * - image: base64 编码的图片数据，或 data URL（如 data:image/png;base64,...）
+     * - aspect_ratio: 宽高比
+     * - image_size: 图片大小（仅 gemini-3-pro-image-preview）
+     * - response_modalities: 响应模态
+     */
+    edit(request: GeminiImageEditRequest): Promise<GeminiImageGenerationResponse>;
+    /**
+     * Chat with images (multi-turn image editing)
+     * 图像多轮对话（用于多轮图片修改）
+     *
+     * 支持通过多轮对话迭代式优化图片
+     *
+     * 使用示例：
+     * ```typescript
+     * // 第一轮：创建初始图片
+     * const firstResponse = await client.chat({
+     *   message: "Create a vibrant infographic about photosynthesis"
+     * });
+     *
+     * // 后续轮次：修改图片（使用返回的 chat_id）
+     * const secondResponse = await client.chat({
+     *   chat_id: firstResponse.chat_id,
+     *   message: "Update this infographic to be in Spanish"
+     * });
+     * ```
+     *
+     * 支持的参数：
+     * - message: 当前轮次的提示词
+     * - chat_id: 聊天 ID（用于多轮对话，首次请求时不需要）
+     * - aspect_ratio: 宽高比
+     * - image_size: 图片大小（仅 gemini-3-pro-image-preview）
+     * - response_modalities: 响应模态
+     */
+    chat(request: GeminiImageChatRequest): Promise<GeminiImageChatResponse>;
 }

package/dist/gemini-image-generation.js

@@ -76,5 +76,122 @@ class GeminiImageGenerationClient {
         (0, log_1.logResponse)(response.status, response.statusText, response.headers, data);
         return data;
     }
+    /**
+     * Edit images
+     * 编辑图像
+     *
+     * 使用文本提示编辑图片，支持添加、移除或修改元素
+     *
+     * 支持的参数：
+     * - image: base64 编码的图片数据，或 data URL（如 data:image/png;base64,...）
+     * - aspect_ratio: 宽高比
+     * - image_size: 图片大小（仅 gemini-3-pro-image-preview）
+     * - response_modalities: 响应模态
+     */
+    async edit(request) {
+        const requestBody = {
+            prompt: request.prompt,
+            image: request.image,
+            model: request.model || "gemini-2.5-flash-image",
+        };
+        // 添加可选参数
+        if (request.aspect_ratio) {
+            requestBody.aspect_ratio = request.aspect_ratio;
+        }
+        if (request.image_size) {
+            requestBody.image_size = request.image_size;
+        }
+        if (request.response_modalities) {
+            requestBody.response_modalities = request.response_modalities;
+        }
+        if (request.user) {
+            requestBody.user = request.user;
+        }
+        if (request.provider) {
+            requestBody.provider = request.provider;
+        }
+        const url = `${config_1.sdkConfig.getServerUrl()}/api/gemini-image-proxy/edit`;
+        (0, log_1.logRequest)("POST", url, this.headers, { ...requestBody, image: "[base64 data]" });
+        const response = await fetch(url, {
+            method: "POST",
+            headers: this.headers,
+            body: JSON.stringify(requestBody),
+        });
+        if (!response.ok) {
+            const errorText = await response.text();
+            (0, log_1.logResponse)(response.status, response.statusText, response.headers, errorText);
+            throw new Error(`Gemini image edit API error: ${response.status} ${errorText}`);
+        }
+        const data = (await response.json());
+        (0, log_1.logResponse)(response.status, response.statusText, response.headers, data);
+        return data;
+    }
+    /**
+     * Chat with images (multi-turn image editing)
+     * 图像多轮对话（用于多轮图片修改）
+     *
+     * 支持通过多轮对话迭代式优化图片
+     *
+     * 使用示例：
+     * ```typescript
+     * // 第一轮：创建初始图片
+     * const firstResponse = await client.chat({
+     *   message: "Create a vibrant infographic about photosynthesis"
+     * });
+     *
+     * // 后续轮次：修改图片（使用返回的 chat_id）
+     * const secondResponse = await client.chat({
+     *   chat_id: firstResponse.chat_id,
+     *   message: "Update this infographic to be in Spanish"
+     * });
+     * ```
+     *
+     * 支持的参数：
+     * - message: 当前轮次的提示词
+     * - chat_id: 聊天 ID（用于多轮对话，首次请求时不需要）
+     * - aspect_ratio: 宽高比
+     * - image_size: 图片大小（仅 gemini-3-pro-image-preview）
+     * - response_modalities: 响应模态
+     */
+    async chat(request) {
+        const requestBody = {
+            message: request.message,
+            model: request.model || "gemini-3-pro-image-preview",
+        };
+        // 添加可选参数
+        if (request.chat_id) {
+            requestBody.chat_id = request.chat_id;
+        }
+        if (request.aspect_ratio) {
+            requestBody.aspect_ratio = request.aspect_ratio;
+        }
+        if (request.image_size) {
+            requestBody.image_size = request.image_size;
+        }
+        if (request.response_modalities) {
+            requestBody.response_modalities = request.response_modalities;
+        }
+        if (request.user) {
+            requestBody.user = request.user;
+        }
+        if (request.provider) {
+            requestBody.provider = request.provider;
+        }
+        const url = `${config_1.sdkConfig.getServerUrl()}/api/gemini-image-proxy/chat`;
+        (0, log_1.logRequest)("POST", url, this.headers, requestBody);
+        const response = await fetch(url, {
+            method: "POST",
+            headers: this.headers,
+            body: JSON.stringify(requestBody),
+        });
+        if (!response.ok) {
+            const errorText = await response.text();
+            (0, log_1.logResponse)(response.status, response.statusText, response.headers, errorText);
+            throw new Error(`Gemini image chat API error: ${response.status} ${errorText}`);
+        }
+        const data = (await response.json());
+        (0, log_1.logResponse)(response.status, response.statusText, response.headers, data);
+        return data;
+    }
 }
 exports.GeminiImageGenerationClient = GeminiImageGenerationClient;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "ai-world-sdk",
-  "version": "1.0.8",
+  "version": "1.0.9",
   "description": "TypeScript SDK for AI World Platform - Chat Models, Image Generation, and Video Generation",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",