thatgfsj-code 3.0.4 → 3.0.7

package/src/config/providers.ts

@@ -1,234 +0,0 @@
-/**
- * Provider definitions and model catalogs
- * Updated with correct API formats and more providers
- */
-
-import type { ProviderConfig, ProviderName, ModelInfo } from './types.js';
-
-// ==================== Provider Definitions ====================
-
-export const PROVIDERS: Record<ProviderName, ProviderConfig> = {
-  siliconflow: {
-    name: 'SiliconFlow (硅基流动)',
-    baseUrl: 'https://api.siliconflow.cn/v1',
-    defaultModel: 'Qwen/Qwen2.5-7B-Instruct',
-    envKeys: ['SILICONFLOW_API_KEY'],
-    format: 'openai',
-  },
-  openai: {
-    name: 'OpenAI',
-    baseUrl: 'https://api.openai.com/v1',
-    defaultModel: 'gpt-4o-mini',
-    envKeys: ['OPENAI_API_KEY'],
-    format: 'openai',
-  },
-  deepseek: {
-    name: 'DeepSeek (深度求索)',
-    baseUrl: 'https://api.deepseek.com',
-    defaultModel: 'deepseek-chat',
-    envKeys: ['DEEPSEEK_API_KEY'],
-    format: 'openai',
-  },
-  kimi: {
-    name: 'Kimi (月之暗面)',
-    baseUrl: 'https://api.moonshot.cn/v1',
-    defaultModel: 'kimi-k2.6',
-    envKeys: ['KIMI_API_KEY', 'MOONSHOT_API_KEY'],
-    format: 'openai',
-  },
-  zhipu: {
-    name: 'Zhipu GLM (智谱)',
-    baseUrl: 'https://open.bigmodel.cn/api/paas/v4',
-    defaultModel: 'glm-4-flash',
-    envKeys: ['ZHIPU_API_KEY', 'GLM_API_KEY'],
-    format: 'openai',
-  },
-  minimax: {
-    name: 'MiniMax',
-    baseUrl: 'https://api.minimax.chat/v1',
-    defaultModel: 'MiniMax-Text-01',
-    envKeys: ['MINIMAX_API_KEY'],
-    format: 'openai',
-  },
-  baichuan: {
-    name: 'Baichuan (百川)',
-    baseUrl: 'https://api.baichuan-ai.com/v1',
-    defaultModel: 'Baichuan4',
-    envKeys: ['BAICHUAN_API_KEY'],
-    format: 'openai',
-  },
-  stepfun: {
-    name: 'Stepfun (阶跃星辰)',
-    baseUrl: 'https://api.stepfun.com/v1',
-    defaultModel: 'step-1-flash',
-    envKeys: ['STEPFUN_API_KEY'],
-    format: 'openai',
-  },
-  doubao: {
-    name: 'Doubao (火山引擎豆包)',
-    baseUrl: 'https://ark.cn-beijing.volces.com/api/v3',
-    defaultModel: 'doubao-1.5-pro-32k',
-    envKeys: ['DOUBAO_API_KEY', 'ARK_API_KEY'],
-    format: 'openai',
-  },
-  anthropic: {
-    name: 'Anthropic (Claude)',
-    baseUrl: 'https://api.anthropic.com/v1',
-    defaultModel: 'claude-sonnet-4-20250514',
-    envKeys: ['ANTHROPIC_API_KEY'],
-    format: 'anthropic',
-  },
-  gemini: {
-    name: 'Google Gemini',
-    baseUrl: 'https://generativelanguage.googleapis.com/v1beta',
-    defaultModel: 'gemini-2.0-flash',
-    envKeys: ['GEMINI_API_KEY'],
-    format: 'gemini',
-  },
-  ollama: {
-    name: 'Ollama (本地)',
-    baseUrl: 'http://localhost:11434/v1',
-    defaultModel: 'llama3.1',
-    envKeys: [],
-    format: 'openai',
-  },
-  ernie: {
-    name: 'ERNIE (百度文心)',
-    baseUrl: 'https://qianfan.baidubce.com/v2',
-    defaultModel: 'ernie-4.5-8k',
-    envKeys: ['ERNIE_API_KEY', 'QIANFAN_API_KEY'],
-    format: 'openai',
-  },
-  custom_openai: {
-    name: '自定义 OpenAI 兼容 (中转站)',
-    baseUrl: '',
-    defaultModel: 'gpt-4o-mini',
-    envKeys: ['CUSTOM_API_KEY'],
-    format: 'openai',
-  },
-  custom_anthropic: {
-    name: '自定义 Anthropic 兼容 (中转站)',
-    baseUrl: '',
-    defaultModel: 'claude-sonnet-4-20250514',
-    envKeys: ['CUSTOM_API_KEY'],
-    format: 'anthropic',
-  },
-};
-
-// ==================== Model Catalogs ====================
-
-export const MODEL_CATALOGS: Record<ProviderName, ModelInfo[]> = {
-  siliconflow: [
-    { id: 'Qwen/Qwen2.5-7B-Instruct', name: 'Qwen2.5-7B', desc: '免费，日常编程' },
-    { id: 'Qwen/Qwen2.5-32B-Instruct', name: 'Qwen2.5-32B', desc: '更强性能' },
-    { id: 'Qwen/Qwen2.5-72B-Instruct', name: 'Qwen2.5-72B', desc: '旗舰模型' },
-    { id: 'Qwen/Qwen3-8B', name: 'Qwen3-8B', desc: '最新 Qwen3' },
-    { id: 'Pro/deepseek-ai/DeepSeek-V3', name: 'DeepSeek-V3', desc: '深度求索' },
-    { id: 'Pro/deepseek-ai/DeepSeek-R1', name: 'DeepSeek-R1', desc: '推理模型' },
-    { id: 'deepseek-ai/DeepSeek-V3', name: 'DeepSeek-V3 (免费)', desc: '免费版' },
-    { id: 'THUDM/glm-4-9b-chat', name: 'GLM-4-9B', desc: '智谱模型' },
-    { id: 'internlm/internlm2_5-7b-chat', name: 'InternLM2.5-7B', desc: '书生模型' },
-  ],
-  openai: [
-    { id: 'gpt-4o-mini', name: 'GPT-4o-mini', desc: '性价比高' },
-    { id: 'gpt-4o', name: 'GPT-4o', desc: '旗舰多模态' },
-    { id: 'gpt-4-turbo', name: 'GPT-4-Turbo', desc: '强性能' },
-    { id: 'o3-mini', name: 'o3-mini', desc: '推理模型' },
-  ],
-  deepseek: [
-    { id: 'deepseek-chat', name: 'DeepSeek-Chat', desc: '通用对话' },
-    { id: 'deepseek-reasoner', name: 'DeepSeek-Reasoner', desc: '推理增强' },
-  ],
-  kimi: [
-    { id: 'kimi-k2.6', name: 'Kimi K2.6', desc: '最新旗舰，支持思考' },
-    { id: 'kimi-k2.5', name: 'Kimi K2.5', desc: '稳定版' },
-    { id: 'moonshot-v1-128k', name: 'Moonshot-128K', desc: '超长上下文' },
-    { id: 'moonshot-v1-32k', name: 'Moonshot-32K', desc: '长上下文' },
-    { id: 'moonshot-v1-8k', name: 'Moonshot-8K', desc: '基础版' },
-  ],
-  zhipu: [
-    { id: 'glm-4-plus', name: 'GLM-4-Plus', desc: '旗舰模型' },
-    { id: 'glm-4-flash', name: 'GLM-4-Flash', desc: '快速免费' },
-    { id: 'glm-4-long', name: 'GLM-4-Long', desc: '超长上下文 1M' },
-    { id: 'glm-4-flashx', name: 'GLM-4-FlashX', desc: '加速版' },
-    { id: 'glm-4-air', name: 'GLM-4-Air', desc: '轻量版' },
-  ],
-  minimax: [
-    { id: 'MiniMax-Text-01', name: 'MiniMax-Text-01', desc: '旗舰模型' },
-    { id: 'abab6.5s-chat', name: 'abab6.5s', desc: '轻量版' },
-  ],
-  baichuan: [
-    { id: 'Baichuan4', name: 'Baichuan4', desc: '最新旗舰' },
-    { id: 'Baichuan3-Turbo', name: 'Baichuan3-Turbo', desc: '快速版' },
-  ],
-  stepfun: [
-    { id: 'step-1-flash', name: 'Step-1-Flash', desc: '快速免费' },
-    { id: 'step-1-8k', name: 'Step-1-8K', desc: '基础版' },
-    { id: 'step-2-16k', name: 'Step-2-16K', desc: '增强版' },
-  ],
-  doubao: [
-    { id: 'doubao-1.5-pro-32k', name: 'Doubao-1.5-Pro-32K', desc: '旗舰版' },
-    { id: 'doubao-1.5-lite-32k', name: 'Doubao-1.5-Lite-32K', desc: '轻量版' },
-    { id: 'doubao-pro-256k', name: 'Doubao-Pro-256K', desc: '超长上下文' },
-  ],
-  anthropic: [
-    { id: 'claude-sonnet-4-20250514', name: 'Claude Sonnet 4', desc: '最新旗舰' },
-    { id: 'claude-3-5-haiku-20241022', name: 'Claude 3.5 Haiku', desc: '快速便宜' },
-    { id: 'claude-3-opus-20240229', name: 'Claude 3 Opus', desc: '最强性能' },
-  ],
-  gemini: [
-    { id: 'gemini-2.0-flash', name: 'Gemini 2.0 Flash', desc: '最新快速' },
-    { id: 'gemini-2.5-pro-preview-05-06', name: 'Gemini 2.5 Pro', desc: '最强推理' },
-    { id: 'gemini-1.5-flash', name: 'Gemini 1.5 Flash', desc: '经典版' },
-  ],
-  ollama: [
-    { id: 'llama3.1', name: 'Llama 3.1', desc: 'Meta 开源' },
-    { id: 'qwen2.5', name: 'Qwen 2.5', desc: '通义千问' },
-    { id: 'deepseek-coder-v2', name: 'DeepSeek Coder V2', desc: '编程专用' },
-    { id: 'codellama', name: 'Code Llama', desc: 'Meta 编程' },
-    { id: 'mistral', name: 'Mistral', desc: '高性能' },
-  ],
-  ernie: [
-    { id: 'ernie-4.5-8k', name: 'ERNIE-4.5-8K', desc: '最新旗舰' },
-    { id: 'ernie-4.0-8k', name: 'ERNIE-4.0-8K', desc: '强性能' },
-    { id: 'ernie-3.5-8k', name: 'ERNIE-3.5-8K', desc: '性价比高' },
-  ],
-  custom_openai: [
-    { id: 'gpt-4o-mini', name: 'GPT-4o-mini', desc: '默认模型，可自定义' },
-    { id: 'gpt-4o', name: 'GPT-4o', desc: '可自定义' },
-    { id: 'claude-sonnet-4-20250514', name: 'Claude Sonnet 4', desc: '可自定义' },
-  ],
-  custom_anthropic: [
-    { id: 'claude-sonnet-4-20250514', name: 'Claude Sonnet 4', desc: '默认模型，可自定义' },
-    { id: 'claude-3-5-haiku-20241022', name: 'Claude 3.5 Haiku', desc: '可自定义' },
-  ],
-};
-
-// ==================== Helpers ====================
-
-export function getModelsForProvider(provider: ProviderName): ModelInfo[] {
-  return MODEL_CATALOGS[provider] || [];
-}
-
-export function getApiKeyFromEnv(provider: ProviderName): string {
-  const providerConfig = PROVIDERS[provider];
-  if (!providerConfig) return '';
-  for (const envKey of providerConfig.envKeys) {
-    if (process.env[envKey]) return process.env[envKey];
-  }
-  return '';
-}
-
-export function listProviders(): Array<{ key: ProviderName; name: string }> {
-  return Object.entries(PROVIDERS).map(([key, val]) => ({
-    key: key as ProviderName,
-    name: val.name,
-  }));
-}
-
-/**
- * Check if a provider name is a custom/relay provider
- */
-export function isCustomProvider(provider: ProviderName): boolean {
-  return provider === 'custom_openai' || provider === 'custom_anthropic';
-}

package/src/config/types.ts

@@ -1,71 +0,0 @@
-/**
- * Configuration types for Thatgfsj Code
- */
-
-export type ProviderName =
-  | 'siliconflow'
-  | 'openai'
-  | 'deepseek'
-  | 'kimi'
-  | 'zhipu'
-  | 'minimax'
-  | 'baichuan'
-  | 'stepfun'
-  | 'doubao'
-  | 'anthropic'
-  | 'gemini'
-  | 'ollama'
-  | 'ernie'
-  | 'custom_openai'
-  | 'custom_anthropic';
-
-export interface ProviderConfig {
-  name: string;
-  baseUrl: string;
-  defaultModel: string;
-  envKeys: string[];
-  format: 'openai' | 'anthropic' | 'gemini';
-}
-
-export interface Config {
-  model: string;
-  apiKey: string;
-  temperature: number;
-  maxTokens: number;
-  contextLength: number;
-  provider: ProviderName;
-  baseUrl?: string;
-  /**
-   * v3.0.0: prompt-cache policy. Default = Anthropic-style explicit
-   * cache_control on (ttl 5m), all other providers fall through to their
-   * built-in automatic caching.
-   *
-   * Note: enabling/disabling only affects Anthropic; other providers do
-   * not expose a programmatic cache toggle, so the flag is purely
-   * advisory there.
-   *
-   * v3.0.3: `ttl` accepts '5m' | '1h' | 'auto'. 'auto' lets the runtime
-   * pick the TTL based on conversation length (decideTTL in smartModel.ts).
-   * The wizard's default is 'auto' so users don't have to choose.
-   */
-  cache?: {
-    enabled: boolean;
-    ttl?: '5m' | '1h' | 'auto';
-    strategy?: 'auto' | 'manual' | 'off';
-  };
-}
-
-export interface AIConfig {
-  model: string;
-  apiKey?: string;
-  temperature?: number;
-  maxTokens?: number;
-  baseUrl?: string;
-  provider?: ProviderName;
-}
-
-export interface ModelInfo {
-  id: string;
-  name: string;
-  desc: string;
-}

package/src/hooks/index.ts

@@ -1,111 +0,0 @@
-/**
- * Hooks System - Event-driven lifecycle hooks
- * Migrated from old src/core/hooks.ts
- */
-
-import type { ChatMessage, ToolCall } from '../types.js';
-
-// ==================== Hook Types ====================
-
-export type HookName =
-  | 'beforeToolCall'
-  | 'afterToolCall'
-  | 'beforeAgentLoop'
-  | 'afterAgentLoop'
-  | 'onError';
-
-export interface HookContext {
-  toolName?: string;
-  toolParams?: Record<string, any>;
-  toolResult?: any;
-  messages?: ChatMessage[];
-  error?: Error;
-  iteration?: number;
-  [key: string]: any;
-}
-
-export type HookFn = (ctx: HookContext) => void | Promise<void>;
-
-interface HookRegistration {
-  name: HookName;
-  fn: HookFn;
-  priority?: number;
-  once?: boolean;
-}
-
-// ==================== Hook Manager ====================
-
-export class HookManager {
-  private hooks: Map<HookName, HookRegistration[]> = new Map();
-
-  constructor() {
-    const allHooks: HookName[] = ['beforeToolCall', 'afterToolCall', 'beforeAgentLoop', 'afterAgentLoop', 'onError'];
-    for (const h of allHooks) {
-      this.hooks.set(h, []);
-    }
-  }
-
-  register(name: HookName, fn: HookFn, priority = 0, once = false): void {
-    const regs = this.hooks.get(name) || [];
-    regs.push({ name, fn, priority, once });
-    regs.sort((a, b) => (b.priority || 0) - (a.priority || 0));
-    this.hooks.set(name, regs);
-  }
-
-  unregister(name: HookName, fn: HookFn): void {
-    const regs = this.hooks.get(name) || [];
-    this.hooks.set(name, regs.filter(r => r.fn !== fn));
-  }
-
-  clear(name?: HookName): void {
-    if (name) {
-      this.hooks.set(name, []);
-    } else {
-      for (const regs of this.hooks.values()) {
-        regs.length = 0;
-      }
-    }
-  }
-
-  async emit(name: HookName, ctx: HookContext): Promise<void> {
-    const regs = this.hooks.get(name) || [];
-    for (const reg of regs) {
-      try {
-        await Promise.resolve(reg.fn(ctx));
-      } catch (error: any) {
-        console.error(`[Hook ${name}] error:`, error.message);
-      }
-      if (reg.once) {
-        this.unregister(name, reg.fn);
-      }
-    }
-  }
-
-  hasHooks(name: HookName): boolean {
-    return (this.hooks.get(name)?.length || 0) > 0;
-  }
-}
-
-// ==================== Built-in Hooks ====================
-
-/**
- * Audit logging hook - logs all tool calls
- */
-export function auditLogHook(ctx: HookContext): void {
-  if (ctx.toolName) {
-    const ts = new Date().toISOString().split('T')[1].split('.')[0];
-    const status = ctx.toolResult?.success ? 'OK' : 'FAIL';
-    console.log(`[AUDIT ${ts}] ${status} | tool=${ctx.toolName} | params=${JSON.stringify(ctx.toolParams || {})}`);
-  }
-}
-
-// ==================== Singleton ====================
-
-let globalHookManager: HookManager | null = null;
-
-export function getHookManager(): HookManager {
-  if (!globalHookManager) {
-    globalHookManager = new HookManager();
-  }
-  return globalHookManager;
-}