autosnippet
Version:
Extract code patterns into a knowledge base for AI coding assistants
184 lines (177 loc) • 6.89 kB
JavaScript
/**
* ChatAgentTasks — Agent 预定义任务方法
*
* 每个任务接收 context 对象: { invokeAgent, aiProvider, container, logger }
* - invokeAgent(toolName, params) — 直接执行工具 handler（纯数据工具）
* - aiProvider — AI Provider 实例
* - container — ServiceContainer
* - logger — Logger 实例
*/
/**
* 任务: 提交前查重 + 质量预评
* 1. check_duplicate → 若发现相似 ≥ 0.7 则建议合并
* 2. 顺便返回质量评估建议
*/
export async function taskCheckAndSubmit(context, { candidate, projectRoot }) {
const { invokeAgent, aiProvider } = context;
// Step 1: 查重
const duplicates = (await invokeAgent('check_duplicate', {
candidate,
projectRoot,
threshold: 0.5,
}));
// Step 2: 如果有高相似度，使用 AI 分析是否真正重复
const highSim = (duplicates.similar || []).filter((d) => d.similarity >= 0.7);
let aiVerdict = null;
if (highSim.length > 0 && aiProvider) {
const verdictPrompt = `以下新候选代码与已有 Recipe 高度相似，请判断是否真正重复。
新候选:
- Title: ${candidate.title || '(未命名)'}
- Code: ${(candidate.code || '').substring(0, 1000)}
相似 Recipe:
${highSim.map((s) => `- ${s.title} (相似度: ${s.similarity})`).join('\n')}
请回答: DUPLICATE（真正重复）/ SIMILAR（相似但不同，建议保留并标注关系）/ UNIQUE（误判，可放心提交）
只回答一个词。`;
try {
const raw = await aiProvider.chat(verdictPrompt, { temperature: 0, maxTokens: 20 });
aiVerdict = (raw || '').trim().toUpperCase().split(/\s/)[0];
}
catch {
/* ignore */
}
}
return {
duplicates: duplicates.similar || [],
highSimilarity: highSim,
aiVerdict,
recommendation: highSim.length === 0
? 'safe_to_submit'
: aiVerdict === 'DUPLICATE'
? 'block_duplicate'
: 'review_suggested',
};
}
/**
* 任务: 批量发现 Recipe 间的知识图谱关系
* 委托给 AgentFactory.discoverRelations (独立 Explore → Synthesize 管线)
*/
export async function taskDiscoverAllRelations(context, { batchSize = 20 } = {}) {
const { container } = context;
const agentFactory = container.get('agentFactory');
// Mock 模式下跳过 AI 关系发现
if (agentFactory.getAiProviderInfo?.()?.name === 'mock') {
return { discovered: 0, message: 'AI Provider 未配置（Mock 模式），跳过关系发现。' };
}
return agentFactory.discoverRelations({ batchSize });
}
/** 任务: 批量 AI 补全候选语义字段 */
export async function taskFullEnrich(context, { status = 'pending', maxCount = 50 } = {}) {
const { invokeAgent, container } = context;
const knowledgeService = container.get('knowledgeService');
const { items = [], data = [] } = await knowledgeService.list({ lifecycle: status }, { page: 1, pageSize: maxCount });
const candidates = items.length > 0 ? items : data;
if (candidates.length === 0) {
return { enriched: 0, message: 'No candidates to enrich' };
}
// 筛选缺失语义字段的候选
const needEnrich = candidates.filter((c) => {
const m = c.metadata || {};
return !m.rationale || !m.knowledgeType || !m.complexity;
});
if (needEnrich.length === 0) {
return { enriched: 0, message: 'All candidates already enriched' };
}
const result = await invokeAgent('enrich_candidate', {
candidateIds: needEnrich.map((c) => c.id).slice(0, 20),
});
return result;
}
/**
* 任务: 批量质量审计全部 Recipe
* 对活跃 Recipe 逐个评分，返回低于阈值的列表
*/
export async function taskQualityAudit(context, { threshold = 0.6, maxCount = 100 } = {}) {
const { invokeAgent, container } = context;
const knowledgeService = container.get('knowledgeService');
const { items = [], data = [] } = await knowledgeService.list({ lifecycle: 'active' }, { page: 1, pageSize: maxCount });
const recipes = items.length > 0 ? items : data;
if (recipes.length === 0) {
return { total: 0, lowQuality: [], message: 'No active recipes' };
}
const lowQuality = [];
const gradeDistribution = { A: 0, B: 0, C: 0, D: 0, F: 0 };
for (const recipe of recipes) {
const scoreResult = (await invokeAgent('quality_score', { recipe }));
if (scoreResult.grade) {
gradeDistribution[scoreResult.grade] =
(gradeDistribution[scoreResult.grade] || 0) + 1;
}
if (scoreResult.score < threshold) {
lowQuality.push({
id: recipe.id,
title: recipe.title,
score: scoreResult.score,
grade: scoreResult.grade,
dimensions: scoreResult.dimensions,
});
}
}
lowQuality.sort((a, b) => a.score - b.score);
return {
total: recipes.length,
threshold,
gradeDistribution,
lowQualityCount: lowQuality.length,
lowQuality,
};
}
/**
* 任务: Guard 完整扫描
* 对代码运行全部 Guard 规则 + 生成修复建议
*/
export async function taskGuardFullScan(context, { code, language, filePath } = {}) {
const { invokeAgent, aiProvider } = context;
if (!code) {
return { error: 'code is required' };
}
// Step 1: 静态检查
const checkResult = (await invokeAgent('guard_check_code', {
code,
language: language || 'unknown',
scope: 'project',
}));
// Step 2: 如果有违规且 AI 可用，生成修复建议
let suggestions = null;
if (checkResult.violationCount > 0 && aiProvider) {
try {
const violationSummary = (checkResult.violations || [])
.slice(0, 5)
.map((v) => `- [${v.severity}] ${v.message || v.ruleName} (line ${v.line || v.matches?.[0]?.line || '?'})`)
.join('\n');
const prompt = `以下代码存在 Guard 规则违规。请为每个违规提供修复建议。
违规列表:
${violationSummary}
代码片段:
\`\`\`${language || ''}
${code.substring(0, 3000)}
\`\`\`
请用 JSON 数组格式返回建议: [{"violation": "...", "suggestion": "...", "fixExample": "..."}]`;
suggestions =
(await aiProvider.chatWithStructuredOutput(prompt, {
openChar: '[',
closeChar: ']',
temperature: 0.3,
})) || [];
}
catch {
/* AI suggestions optional */
}
}
return {
filePath: filePath || '(inline)',
language,
violationCount: checkResult.violationCount,
violations: checkResult.violations,
suggestions,
};
}