@next-open-ai/openclawx 0.8.40 → 0.8.58

package/dist/server/config/local-models.service.js

@@ -0,0 +1,243 @@
+var __decorate = (this && this.__decorate) || function (decorators, target, key, desc) {
+    var c = arguments.length, r = c < 3 ? target : desc === null ? desc = Object.getOwnPropertyDescriptor(target, key) : desc, d;
+    if (typeof Reflect === "object" && typeof Reflect.decorate === "function") r = Reflect.decorate(decorators, target, key, desc);
+    else for (var i = decorators.length - 1; i >= 0; i--) if (d = decorators[i]) r = (c < 3 ? d(r) : c > 3 ? d(target, key, r) : d(target, key)) || r;
+    return c > 3 && r && Object.defineProperty(target, key, r), r;
+};
+var __metadata = (this && this.__metadata) || function (k, v) {
+    if (typeof Reflect === "object" && typeof Reflect.metadata === "function") return Reflect.metadata(k, v);
+};
+/**
+ * 本地 GGUF 模型管理服务。
+ * 负责列出、下载（通过 node-llama-cpp resolveModelFile）、删除本地缓存的 GGUF 模型文件。
+ * 推荐列表从 presets/recommended-local-models.json 加载，已安装的与推荐使用同一套展示名称且已安装的不再出现在「备下载」列表。
+ * 模型缓存目录：~/.openbot/.cached_models/
+ */
+import { Injectable } from '@nestjs/common';
+import { readdir, stat, unlink, readFile } from 'node:fs/promises';
+import { join, basename } from 'node:path';
+import { existsSync, mkdirSync } from 'node:fs';
+import { modelUriToFilename, modelUriBasename, LOCAL_LLM_CACHE_DIR } from '../../core/local-llm-server/model-resolve.js';
+/** 根据文件名推断模型类型 */
+function inferModelType(filename) {
+    const lower = filename.toLowerCase();
+    if (lower.includes('embed') || lower.includes('bge') || lower.includes('e5-'))
+        return 'embedding';
+    return 'llm';
+}
+const PRESET_FILENAME = 'recommended-local-models.json';
+const DEFAULT_RECOMMENDED = [
+    { id: 'hf:Qwen/Qwen3-4B-GGUF/Qwen3-4B-Q4_K_M.gguf', name: 'Qwen3 4B Q4_K_M', type: 'llm', sizeHint: '~2.5GB' },
+    { id: 'hf:Qwen/Qwen3-7B-GGUF/Qwen3-7B-Q4_K_M.gguf', name: 'Qwen3 7B Q4_K_M', type: 'llm', sizeHint: '~4.5GB' },
+    { id: 'hf:Qwen/Qwen3-14B-GGUF/Qwen3-14B-Q4_K_M.gguf', name: 'Qwen3 14B Q4_K_M', type: 'llm', sizeHint: '~8.5GB' },
+    { id: 'hf:ggml-org/embeddinggemma-300M-GGUF/embeddinggemma-300M-Q8_0.gguf', name: 'EmbeddingGemma 300M Q8 (768维)', type: 'embedding', sizeHint: '~300MB' },
+    { id: 'hf:gpustack/bge-m3-GGUF/bge-m3-Q8_0.gguf', name: 'BGE-M3 Q8 多语言 (1024维)', type: 'embedding', sizeHint: '~1.2GB' },
+    { id: 'hf:mixedbread-ai/mxbai-embed-large-v1-GGUF/mxbai-embed-large-v1-f16.gguf', name: 'MxBai Embed Large v1 (1024维)', type: 'embedding', sizeHint: '~670MB' },
+];
+function getPresetsDir() {
+    return process.env.OPENBOT_PRESETS_DIR || join(process.cwd(), 'presets');
+}
+/** 推荐项是否对应已安装文件（精确预测名 或 安装文件名以 uri 末尾文件名结尾，兼容不同 node-llama-cpp 命名） */
+function recommendedMatchesInstalled(rec, installedFilenames) {
+    const predicted = modelUriToFilename(rec.id);
+    if (installedFilenames.has(predicted))
+        return true;
+    const suffix = modelUriBasename(rec.id);
+    return Array.from(installedFilenames).some((f) => f.endsWith(suffix) || f === suffix);
+}
+/** 已安装文件名在预设中对应的展示名，若无匹配则返回 filename */
+function displayNameForFilename(filename, recommended) {
+    for (const rec of recommended) {
+        const predicted = modelUriToFilename(rec.id);
+        if (filename === predicted)
+            return rec.name;
+        const uriBase = modelUriBasename(rec.id);
+        if (uriBase && filename.endsWith(uriBase))
+            return rec.name;
+    }
+    return filename;
+}
+let LocalModelsService = class LocalModelsService {
+    cacheDir;
+    recommendedCache = null;
+    /** 正在下载的任务：modelUri → 进度 */
+    downloadingMap = new Map();
+    constructor() {
+        this.cacheDir = LOCAL_LLM_CACHE_DIR;
+        if (!existsSync(this.cacheDir)) {
+            mkdirSync(this.cacheDir, { recursive: true });
+        }
+    }
+    /** 从 presets/recommended-local-models.json 加载推荐列表，失败则用内置默认 */
+    async getRecommendedModelsFromPreset() {
+        if (this.recommendedCache)
+            return this.recommendedCache;
+        const presetPath = join(getPresetsDir(), PRESET_FILENAME);
+        try {
+            if (existsSync(presetPath)) {
+                const raw = await readFile(presetPath, 'utf-8');
+                const data = JSON.parse(raw);
+                if (Array.isArray(data.models) && data.models.length > 0) {
+                    this.recommendedCache = data.models;
+                    return this.recommendedCache;
+                }
+            }
+        }
+        catch {
+            // ignore, fallback to default
+        }
+        this.recommendedCache = DEFAULT_RECOMMENDED;
+        return this.recommendedCache;
+    }
+    /** 列出本地已缓存的 GGUF 模型文件，displayName 与推荐列表一致（来自 preset 匹配） */
+    async listModels() {
+        try {
+            const recommended = await this.getRecommendedModelsFromPreset();
+            const files = await readdir(this.cacheDir);
+            const ggufFiles = files.filter((f) => f.endsWith('.gguf'));
+            const infos = await Promise.all(ggufFiles.map(async (f) => {
+                const filePath = join(this.cacheDir, f);
+                const s = await stat(filePath);
+                return {
+                    filename: f,
+                    size: s.size,
+                    updatedAt: s.mtime.toISOString(),
+                    inferredType: inferModelType(f),
+                    displayName: displayNameForFilename(f, recommended),
+                };
+            }));
+            return infos.sort((a, b) => b.updatedAt.localeCompare(a.updatedAt));
+        }
+        catch {
+            return [];
+        }
+    }
+    /** 删除本地缓存的 GGUF 模型文件 */
+    async deleteModel(filename) {
+        // 安全检查：只允许删除 .gguf 文件，且不含路径分隔符
+        if (!filename.endsWith('.gguf') || filename.includes('/') || filename.includes('\\')) {
+            throw new Error('非法文件名');
+        }
+        const filePath = join(this.cacheDir, filename);
+        if (!existsSync(filePath))
+            throw new Error(`文件不存在: ${filename}`);
+        await unlink(filePath);
+    }
+    /** 获取推荐模型列表（来自 preset，与已安装展示名称一致） */
+    async getRecommendedModels() {
+        return this.getRecommendedModelsFromPreset();
+    }
+    /** 仅返回尚未安装的推荐模型；已安装的视为「推荐列表中的一员」不再出现在备下载区 */
+    async getRecommendedToDownload() {
+        const [recommended, installed] = await Promise.all([
+            this.getRecommendedModelsFromPreset(),
+            this.listModels(),
+        ]);
+        const installedFilenames = new Set(installed.map((m) => m.filename));
+        return recommended.filter((rec) => !recommendedMatchesInstalled(rec, installedFilenames));
+    }
+    /** 检查指定模型（uri 或文件名）是否已在缓存目录存在 */
+    isModelFilePresent(modelIdOrUri) {
+        const filename = modelUriToFilename(modelIdOrUri);
+        if (!filename || !filename.endsWith('.gguf'))
+            return false;
+        return existsSync(join(this.cacheDir, filename));
+    }
+    /** 正在下载任务的 AbortController，用于取消 */
+    abortControllerMap = new Map();
+    /** 获取某个 modelUri 的下载进度（不存在则表示未在下载） */
+    getDownloadProgress(modelUri) {
+        return this.downloadingMap.get(modelUri) ?? null;
+    }
+    /**
+     * 取消指定 modelUri 的下载（若正在下载）。
+     */
+    cancelDownload(modelUri) {
+        const controller = this.abortControllerMap.get(modelUri);
+        if (controller) {
+            controller.abort();
+            this.downloadingMap.set(modelUri, { status: '已取消' });
+            this.abortControllerMap.delete(modelUri);
+            setTimeout(() => this.downloadingMap.delete(modelUri), 3000);
+        }
+    }
+    /**
+     * 返回所有推荐模型及是否已安装，用于前端展示「已下载」或下载按钮。
+     */
+    async getRecommendedWithStatus() {
+        const [recommended, installed] = await Promise.all([
+            this.getRecommendedModelsFromPreset(),
+            this.listModels(),
+        ]);
+        const installedFilenames = new Set(installed.map((m) => m.filename));
+        return recommended.map((rec) => ({
+            ...rec,
+            isInstalled: recommendedMatchesInstalled(rec, installedFilenames),
+        }));
+    }
+    /**
+     * 后台下载模型（通过 node-llama-cpp resolveModelFile）。
+     * useMirror=true 使用国内镜像 hf-mirror.com，false 使用官方 huggingface.co。
+     * 返回后立即响应，进度通过 getDownloadProgress 轮询获取。
+     */
+    async startDownload(modelUri, options) {
+        if (this.downloadingMap.has(modelUri)) {
+            return { filename: modelUriToFilename(modelUri) };
+        }
+        const controller = new AbortController();
+        this.abortControllerMap.set(modelUri, controller);
+        this.downloadingMap.set(modelUri, { status: '准备下载...' });
+        this.runDownload(modelUri, controller.signal, options?.useMirror === true).catch((e) => {
+            if (e?.name === 'AbortError' || (typeof e?.message === 'string' && e.message.includes('abort'))) {
+                this.downloadingMap.set(modelUri, { status: '已取消' });
+            }
+            else {
+                const msg = e instanceof Error ? e.message : String(e);
+                let errorMsg = `失败: ${msg}`;
+                if (msg.includes('401') || msg.includes('Unauthorized')) {
+                    errorMsg = '下载失败: 需要 Hugging Face Token。请设置环境变量 HF_TOKEN 或 HUGGING_FACE_TOKEN';
+                }
+                this.downloadingMap.set(modelUri, { status: errorMsg });
+            }
+            this.abortControllerMap.delete(modelUri);
+            setTimeout(() => this.downloadingMap.delete(modelUri), 5000);
+        });
+        return { filename: modelUriToFilename(modelUri) };
+    }
+    async runDownload(modelUri, signal, useMirror) {
+        const { resolveModelFile } = await import('node-llama-cpp');
+        this.downloadingMap.set(modelUri, { status: '解析模型地址...' });
+        const hfToken = process.env.HF_TOKEN || process.env.HUGGING_FACE_TOKEN;
+        const options = {
+            directory: this.cacheDir,
+            onProgress: ({ downloadedSize, totalSize }) => {
+                if (signal?.aborted)
+                    return;
+                const percent = totalSize ? Math.round((downloadedSize / totalSize) * 100) : 0;
+                this.downloadingMap.set(modelUri, {
+                    status: '下载中',
+                    completed: downloadedSize,
+                    total: totalSize,
+                    percent,
+                });
+            },
+            signal,
+            endpoints: {
+                huggingFace: useMirror ? 'https://hf-mirror.com/' : 'https://huggingface.co/',
+            },
+        };
+        if (hfToken) {
+            options.headers = { Authorization: `Bearer ${hfToken}` };
+        }
+        const resolved = await resolveModelFile(modelUri, options);
+        const filename = basename(resolved);
+        this.downloadingMap.set(modelUri, { status: `完成: ${filename}`, percent: 100 });
+        this.abortControllerMap.delete(modelUri);
+        setTimeout(() => this.downloadingMap.delete(modelUri), 5000);
+    }
+};
+LocalModelsService = __decorate([
+    Injectable(),
+    __metadata("design:paramtypes", [])
+], LocalModelsService);
+export { LocalModelsService };

package/package.json

@@ -3,7 +3,7 @@
     "publishConfig": {
         "access": "public"
     },
-    "version": "0.8.40",
+    "version": "0.8.58",
     "description": "OpenClawX - A professional desktop application for managing and executing AI agents with real-time chat, session management, and skills browsing.",
     "type": "module",
     "main": "dist/index.js",

package/presets/preset-agents.json

@@ -68,7 +68,7 @@
       "useLongMemory": true,
       "mcpServers": {
         "akshare-tools": {
-          "command": "npx",
+          "command": "uvx",
           "args": ["-y", "akshare-tools"],
           "env": {}
         }
@@ -107,7 +107,11 @@
         "影刀 RPA MCP Server": {
           "command": "npx",
           "args": ["-y", "yingdao-mcp-server"],
-          "env": {}
+          "env": {
+            "RPA_MODEL": "qwen3-4b-q4_k_m",
+            "SHADOWBOT_PATH": "/Applications/影刀.app",
+            "USER_FOLDER": "/Users/ctrip/Library/Application Support/Shadowbot/users/718771925823332354"
+      }
         }
       }
     },

package/presets/preset-config.json

@@ -1,11 +1,29 @@
 {
     "presetVersion": "1.0",
     "config": {
-        "defaultProvider": "deepseek",
-        "defaultModel": "deepseek-chat",
+        "defaultProvider": "local",
+        "defaultModel": "hf_Qwen_Qwen3-4B-GGUF_Qwen3-4B-Q4_K_M.gguf",
+        "defaultModelItemCode": "local-qwen3-4b",
         "defaultAgentId": "default",
-        "maxAgentSessions": 50,
-        "providers": {},
-        "configuredModels": []
+        "maxAgentSessions": 5,
+        "providers": {
+            "local": { "baseUrl": "http://127.0.0.1:11435/v1" }
+        },
+        "configuredModels": [
+            {
+                "provider": "local",
+                "modelId": "hf_Qwen_Qwen3-4B-GGUF_Qwen3-4B-Q4_K_M.gguf",
+                "type": "llm",
+                "alias": "Qwen3 4B Q4_K_M",
+                "modelItemCode": "local-qwen3-4b"
+            },
+            {
+                "provider": "local",
+                "modelId": "hf_ggml-org_embeddinggemma-300M-GGUF_embeddinggemma-300M-Q8_0.gguf",
+                "type": "embedding",
+                "alias": "EmbeddingGemma 300M Q8 (768维)",
+                "modelItemCode": "local-embeddinggemma-300m"
+            }
+        ]
     }
-}
+}

package/presets/recommended-local-models.json

@@ -0,0 +1,42 @@
+{
+  "presetVersion": "1.0",
+  "description": "推荐本地 GGUF 模型列表，用于「模型管理」备下载与已安装展示。版本升级时可在此增删改，以支持更合适的模型。",
+  "models": [
+    {
+      "id": "hf:Qwen/Qwen3-4B-GGUF/Qwen3-4B-Q4_K_M.gguf",
+      "name": "Qwen3 4B Q4_K_M",
+      "type": "llm",
+      "sizeHint": "~2.5GB"
+    },
+    {
+      "id": "hf:Qwen/Qwen3-7B-GGUF/Qwen3-7B-Q4_K_M.gguf",
+      "name": "Qwen3 7B Q4_K_M",
+      "type": "llm",
+      "sizeHint": "~4.5GB"
+    },
+    {
+      "id": "hf:Qwen/Qwen3-14B-GGUF/Qwen3-14B-Q4_K_M.gguf",
+      "name": "Qwen3 14B Q4_K_M",
+      "type": "llm",
+      "sizeHint": "~8.5GB"
+    },
+    {
+      "id": "hf:ggml-org/embeddinggemma-300M-GGUF/embeddinggemma-300M-Q8_0.gguf",
+      "name": "EmbeddingGemma 300M Q8 (768维)",
+      "type": "embedding",
+      "sizeHint": "~300MB"
+    },
+    {
+      "id": "hf:gpustack/bge-m3-GGUF/bge-m3-Q8_0.gguf",
+      "name": "BGE-M3 Q8 多语言 (1024维)",
+      "type": "embedding",
+      "sizeHint": "~1.2GB"
+    },
+    {
+      "id": "hf:mixedbread-ai/mxbai-embed-large-v1-GGUF/mxbai-embed-large-v1-f16.gguf",
+      "name": "MxBai Embed Large v1 (1024维)",
+      "type": "embedding",
+      "sizeHint": "~670MB"
+    }
+  ]
+}