@promptx/cli
Version:
DPML-powered AI prompt framework - Revolutionary AI-First CLI system based on Deepractice Prompt Markup Language. Build sophisticated AI agents with structured prompts, memory systems, and execution frameworks.
298 lines (257 loc) • 9.11 kB
JavaScript
const BasePouchCommand = require('../BasePouchCommand')
const { getGlobalResourceManager } = require('../../resource')
const { CognitionManager } = require('../../cognition/CognitionManager')
const logger = require('../../../utils/logger')
/**
* 思考锦囊命令 - 基于认知心理学的思维链式推理
* 使用 CognitionManager 进行递归深化的思考过程
*/
class ThinkCommand extends BasePouchCommand {
constructor () {
super()
this.resourceManager = getGlobalResourceManager()
this.cognitionManager = new CognitionManager(this.resourceManager)
}
getPurpose () {
return 'AI主动深度思考，通过认知循环生成洞察和结论'
}
async getContent (args) {
// 解析参数：role、thought对象
const { role, thought } = this.parseArgs(args)
if (!role || !thought) {
return this.getUsageHelp()
}
try {
logger.step('🤔 [ThinkCommand] 开始思考流程')
logger.info(`🧠 [ThinkCommand] 角色: ${role}, 模式: ${thought.thinkingPattern || '未指定'}`)
// 处理 spreadActivationCues：如果是字符串，转换为数组
if (thought.spreadActivationCues && typeof thought.spreadActivationCues === 'string') {
thought.spreadActivationCues = thought.spreadActivationCues.split(' ').filter(cue => cue.trim() !== '');
}
// 验证必需字段
if (!thought.goalEngram) {
throw new Error('Thought 必须包含 goalEngram')
}
if (!thought.thinkingPattern) {
throw new Error('Thought 必须包含 thinkingPattern')
}
if (!thought.spreadActivationCues || thought.spreadActivationCues.length === 0) {
throw new Error('Thought 必须包含 spreadActivationCues')
}
// 使用 CognitionManager 进行思考
const prompt = await this.cognitionManager.think(role, thought)
logger.success('✅ [ThinkCommand] 思考指导生成完成')
return this.formatThinkResponse(thought, prompt, role)
} catch (error) {
logger.error(`❌ [ThinkCommand] 思考失败: ${error.message}`)
logger.error(`🐛 [ThinkCommand] 错误堆栈:\n${error.stack}`)
return `❌ 思考失败：${error.message}
📋 **错误堆栈**：
\`\`\`
${error.stack}
\`\`\`
💡 **可能的原因**：
- 角色ID不正确
- Thought 对象格式错误
- 缺少必需的字段（goalEngram、thinkingPattern、spreadActivationCues）
- 思维模式不存在
🔧 **建议操作**：
1. 确保 Thought 包含所有必需字段
2. 检查角色是否已激活
3. 验证思维模式名称是否正确`
}
}
/**
* 解析命令行参数
*/
parseArgs(args) {
let role = ''
let thought = null
// 第一个参数是role
if (args.length > 0) {
role = args[0]
}
// 第二个参数是JSON格式的thought对象
if (args.length > 1) {
try {
thought = JSON.parse(args[1])
if (typeof thought !== 'object') {
throw new Error('thought必须是对象格式')
}
} catch (error) {
logger.error(`❌ [ThinkCommand] 解析thought参数失败: ${error.message}`)
thought = null
}
}
return { role, thought }
}
/**
* 格式化思考响应
*/
formatThinkResponse (thought, prompt, role) {
const hasInsights = thought.insightEngrams && thought.insightEngrams.length > 0
const hasConclusion = !!thought.conclusionEngram
const hasConfidence = thought.confidence !== undefined
let status = '初始思考'
if (hasConfidence) {
status = '完整思考'
} else if (hasConclusion) {
status = '形成结论'
} else if (hasInsights) {
status = '产生洞察'
}
return `🧠 思考指导已生成
## 📊 当前思考状态
- **角色**: ${role}
- **状态**: ${status}
- **目标**: ${thought.goalEngram.content}
- **思维模式**: ${thought.thinkingPattern}
- **激活线索**: ${thought.spreadActivationCues.join(', ')}
## 💭 生成的思考指导
${prompt}
## 📊 当前进展
${hasInsights ? `- **洞察数量**: ${thought.insightEngrams.length}` : '- **洞察**: 尚未生成'}
${hasConclusion ? `- **已形成结论**: ${thought.conclusionEngram.content}` : '- **结论**: 尚未形成'}
${hasConfidence ? `- **置信度**: ${thought.confidence}` : '- **置信度**: 尚未评估'}
## 🔄 思考深化建议
${this.getDeepingAdvice(thought)}`
}
/**
* 获取思考深化建议
*/
getDeepingAdvice(thought) {
const hasInsights = thought.insightEngrams && thought.insightEngrams.length > 0
const hasConclusion = !!thought.conclusionEngram
const hasConfidence = thought.confidence !== undefined
if (!hasInsights) {
return '- 基于检索到的记忆，生成关键洞察'
} else if (!hasConclusion) {
return '- 综合洞察形成明确结论'
} else if (!hasConfidence) {
return '- 评估结论的置信度'
} else {
return '- 思考已完整，可以开始新的思考目标'
}
}
/**
* 获取使用帮助
*/
getUsageHelp () {
return `🤔 **Think锦囊 - AI深度思考系统**
## 📖 基本用法
think 角色ID '{"goalEngram": {...}, "thinkingPattern": "...", "spreadActivationCues": [...]}'
## 🎯 必填参数
- **角色ID**: 进行思考的角色ID
- **thought对象**: JSON格式的Thought对象，必须包含：
- **goalEngram**: 思考目标
- **thinkingPattern**: 思维模式
- **spreadActivationCues**: 激活线索
## 💭 Thought 结构
\`\`\`json
{
"goalEngram": {
"content": "推理天空呈现蓝色的光学原理",
"schema": "自然现象\\n 光学现象\\n 大气散射"
},
"thinkingPattern": "reasoning",
"spreadActivationCues": ["光学", "大气", "散射", "颜色"],
"insightEngrams": [...], // 可选
"conclusionEngram": {...}, // 可选
"confidence": 0.95 // 可选
}
\`\`\`
## 📋 使用示例
\`\`\`bash
# 第一次思考
think scientist '{"goalEngram": {"content": "推理天空蓝色原理", "schema": "物理学\\n 光学"}, "thinkingPattern": "reasoning", "spreadActivationCues": ["光学", "大气"]}'
# 深入思考
think scientist '{"goalEngram": {...}, "thinkingPattern": "reasoning", "spreadActivationCues": [...], "insightEngrams": [...]}'
# 使用创造性思维
think writer '{"goalEngram": {...}, "thinkingPattern": "creative", "spreadActivationCues": [...]}'
\`\`\`
## 🧠 思维模式
- **reasoning**: 推理思维（逻辑分析）
- **creative**: 创造性思维 [未实现]
- **critical**: 批判性思维 [未实现]
- **systematic**: 系统性思维 [未实现]
- **narrative**: 叙事思维 [未实现]
- **intuitive**: 直觉思维 [未实现]
- **analytical**: 分析思维 [未实现]
- **experiential**: 经验思维 [未实现]
## 🔍 配套工具
- **激活角色**: action 工具激活角色并启动语义网络
- **检索记忆**: recall 工具为思考提供记忆支持
- **保存洞察**: remember 工具保存重要的思考成果`
}
/**
* 获取PATEOAS导航信息
*/
getPATEOAS (args) {
const hasThought = args.length >= 2
if (!hasThought) {
return {
currentState: 'think_awaiting_input',
availableTransitions: ['action', 'welcome'],
nextActions: [
{
name: '激活角色',
description: '选择并激活思考角色',
method: 'MCP PromptX action 工具',
priority: 'high'
},
{
name: '查看角色',
description: '查看可用角色列表',
method: 'MCP PromptX welcome 工具',
priority: 'medium'
}
]
}
}
return {
currentState: 'thinking_in_progress',
availableTransitions: ['think', 'remember', 'recall'],
nextActions: [
{
name: '继续思考',
description: '基于生成的prompt继续深化思考',
method: 'MCP PromptX think 工具',
priority: 'high'
},
{
name: '保存洞察',
description: '将重要洞察保存为记忆',
method: 'MCP PromptX remember 工具',
priority: 'medium'
},
{
name: '检索记忆',
description: '检索相关记忆支持思考',
method: 'MCP PromptX recall 工具',
priority: 'medium'
}
],
metadata: {
thinkingRole: args[0],
thinkingDepth: this.getThinkingDepth(args[1]),
timestamp: new Date().toISOString(),
systemVersion: '锦囊串联状态机 v1.0'
}
}
}
/**
* 分析思考深度
*/
getThinkingDepth(thoughtStr) {
try {
const thought = JSON.parse(thoughtStr)
if (thought.confidence !== undefined) return 'complete'
if (thought.conclusionEngram) return 'conclusion'
if (thought.insightEngrams && thought.insightEngrams.length > 0) return 'insights'
return 'initial'
} catch {
return 'unknown'
}
}
}
module.exports = ThinkCommand